Upload
others
View
19
Download
0
Embed Size (px)
Citation preview
ASSP CMOS
SPARClite Series 32-Bit RISC Embedded Processor
MB86831
Package
• 176-pin, Plastic SQFP• FPT-176P-M01
Features
• 66 or 80 MHz CPU with on-chip clock multiplier
• SPARC high performance RISC architecture
• 8 window, 136 word register file
• 16 address spaces, 256 Mbyte each
• Harvard-style separate on-chip instruction and data buses
• 4 Kbyte 2-way set-associative instruction cache
• 2 Kbyte 2-way set-associative data cache
• Flexible locking mechanism for data and instruction cache entries
• Option to force non-cached operation for memory areas selected by the programmable chip selects; also qualification on a cycle-by-cycle basis using the Non-Cache Pin
• Four-level buffered writes and one-level instruction pre-fetching
• CPU and core logic up to 5 times the frequency of the bus interface unit using on-chip clock multiplier (33 MHz maximum BIU frequency.)
• Bus interface support for 8-, 16-, or 32-bit wide memory
• Support for burst mode cache fills
• DRAM controller with fast page or burst-mode EDO DRAM support
• Interrupt controller with fast response time and programmable priority
• Burst mode ROM support
• Power-saving sleep mode
• Programmable address decoder and wait-state generator
• Single vector trapping
• 0.35 micron gate, 2-level metal CMOS technology, 3.3V internal with 3.3 or 5V I/O
MB86831
Fujitsu Microelectronics, Inc.
3
Table of Contents
General Discussion ...............................................................................................................................................................5
Pin Configuration.......................................................................................................................................................... 5Block Diagram of MB8683..............................................................................................................................................6Ordering Code ...............................................................................................................................................................6Pin Assignment—176-pin SQFP ......................................................................................................................................7Signal Descriptions ........................................................................................................................................................8Pin Status Description .................................................................................................................................................. 12Access Type .................................................................................................................................................................. 12
Overview ........................................................................................................................................................................... 13
Key Features ................................................................................................................................................................ 13Sleep Mode ........................................................................................................................................................... 14
MB86831 Instruction Set .............................................................................................................................................. 14CPU .................................................................................................................................................................... 15
Address Space .............................................................................................................................................................. 15Registers ..................................................................................................................................................................... 15Instruction Sets ............................................................................................................................................................ 16Interrupt .................................................................................................................................................................... 16Cache ......................................................................................................................................................................... 17Bus Interface ............................................................................................................................................................... 18Interrupt Controller ..................................................................................................................................................... 18Interrupt Trigger Modes ................................................................................................................................................ 18
Request Sense Register (Read only) ........................................................................................................................... 19IRC Mode Register ................................................................................................................................................ 19
Interrupt Controller Modes ........................................................................................................................................... 19Interrupt Controller Operation ............................................................................................................................... 20
Clock Generator .......................................................................................................................................................... 20Clock Multiplication Factor .......................................................................................................................................... 20DRAM Controller ........................................................................................................................................................ 21Programmable Chip Select ............................................................................................................................................. 21Use of Chip Select ........................................................................................................................................................ 21
CS0 ..................................................................................................................................................................... 22CS1 and CS2 ......................................................................................................................................................... 22CS3 ..................................................................................................................................................................... 22CS4 ..................................................................................................................................................................... 22CS5 ..................................................................................................................................................................... 22
SPARClite Series 32-Bit RISC Embedded Processor
4
Fujitsu Microelectronics, Inc.
Idle Cycle Insertion Function .........................................................................................................................................22IDLEEN Pin Tied Low ...........................................................................................................................................22IDLEEN Pin Tied High ..........................................................................................................................................22
Idle Cycle Insertion Function .........................................................................................................................................23MB86831 Control and Status Registers—Read/Write .......................................................................................................24MB86831 Memory Mapped Control Registers—Read/Write ..............................................................................................25
Bus Operation .....................................................................................................................................................................31
Operation of the BIU ....................................................................................................................................................31Exception Handling ...............................................................................................................................................31Bus Cycles .............................................................................................................................................................31Load ....................................................................................................................................................................31Load (32-bit bus width) ..........................................................................................................................................31Load (16-bit bus width) ..........................................................................................................................................32Load (8-bit wide bus) .............................................................................................................................................32Load with Exception ..............................................................................................................................................32Store ....................................................................................................................................................................32Store (32-bit bus width) ..........................................................................................................................................32Store (16-bit wide bus) ...........................................................................................................................................32Store (8-bit wide bus) .............................................................................................................................................32Store with Exception ..............................................................................................................................................32Atomic Load Store .................................................................................................................................................32External Bus Request and Grant ...............................................................................................................................328- and 16-Bit Bus Modes .........................................................................................................................................33
Bus Width Control of CS0 .............................................................................................................................................33Bus Width Control Bits of CS1 to CS5 .............................................................................................................................33
Burst Mode Transactions .........................................................................................................................................33ADR<3:2> Sequence in Burst Mode ...............................................................................................................................33
Selection of Hyperpage Mode (EDO Mode) ...............................................................................................................33Basic DRAM Access Timing ............................................................................................................................................34DRAM Access Timing ....................................................................................................................................................34
Electrical Characteristics ......................................................................................................................................................46
Absolute Maximum Ratings ............................................................................................................................................46Recommended Operating Conditions ...............................................................................................................................46DC Characteristics ........................................................................................................................................................47AC Characteristics .......................................................................................................................................................48
Exterior Package Drawing ...................................................................................................................................................53
MB86831
Fujitsu Microelectronics, Inc.
5
General Discussion
The MB86831 is a member of the MB8683X series of RISC processors which offers high performance and high integration for a wide range of embedded applications. The processor is based on the SPARC architecture and is upward code-compatible with previous implementations. At 66 and 80 MHz, the processor runs at up to 66 and 80 MIPS, respectively.
The MB86831 is housed in a low-profile 176-pin plastic package. The on-chip data and instruction caches help decouple the processor from external memory latency. Separate on-chip instruction and data paths provide a high-bandwidth interface between the integer unit (IU) and caches.
Pin Configuration
For maximum performance with a minimum of glue logic, the MB86831 includes: programmable chip select outputs and wait state generation, built-in support for page-mode DRAM, EDO DRAM, and page-mode EEPROM, and support for 8 and 16 and 32-bit memory. These features combine to give the MB86831 high integration and high performance, with the flexibility and efficiency to make it the ideal choice for a wide variety of low cost, embedded systems.
89
1 44
133
176
88
45
TOP VIEW
132
SPARClite Series 32-Bit RISC Embedded Processor
6
Fujitsu Microelectronics, Inc.
Block Diagram of MB86831
Table 1. Ordering Code
Note: The ordering code for production level product. Early shipments of this device may be marked with “ES” to indicate that the part is not yet at full production status.
Clock Frequency (MHz) Ordering Code Package Type
66 MB86831-66PFV-G Plastic SQFP 176
80 MB86831-80PFV-G Plastic SQFP 176
ADR/CONTROL
INTERFACE TOMB86831 BUS
BIU
DCACHE2 KBYTESIU
ICACHE 4 KBYTES
ID
DD
32 32 32 32
3232
32DATA
32DATA
32
ROM, EEPROM CONTROLLER,
ADDRESS DECODEWAIT STATE GENERATOR
ADR/CONTROL
PROGRAMMABLEINTERRUPT
CONTROLLER
8
IRL/IRQ
DRAM/EDO DRAM CONTROLLER
32DATA
32
MB86831
Fujitsu Microelectronics, Inc.
7
* These inputs have a 50K
Ω
internal pullup resistor.
Table 2. Pin Assignment – 76-pin SQFP
Pin No. I/O Pin Name Pin
No. I/O Pin Name Pin No. I/O Pin Name Pin
No. I/O Pin Name
1 VDD 45 VSS 89 VDD 133 VSS
2 I/O D31 46 O DWE3 90 O BE3 / ADR0 134 I/O ASI3
3 I/O D30 47 O DWE2 91 I/O BE2 / ADR1 135 I/O ASI2
4 I/O D29 48 O DWE1 92 O BE1 136 I/O ASI1
5 I/O D28 49 O DWE0 93 O BE0 137 I/O ASI0
6 VSS 50 VSS 94 VSS 138 VSS
7 I BMODE16 51 IO_VDD 95 I NONCACHE 139 IO_VDD
8 I/O D27 52 O RAS0 96 N/A N.C. 140 I IRL3 / IRQ15
9 I/O D26 53 O RAS1 97 N/A N.C. 141 I IRL2 / IRQ14
10 I/O D25 54 O RAS2 98 I/O ADR2 142 I IRL1 / IRQ13
11 I/O D24 55 O RAS3 99 I/O ADR3 143 I IRL0 / IRQ12
12 IO_VDD 56 VDD 100 IO_VDD 144 VDD
13 I/O D23 57 O CAS0 101 I/O ADR4 145 I FLOAT
14 I/O D22 58 O CAS1 102 I/O ADR5 146 O PDOWN
15 I/O D21 59 O CAS2 103 I/O ADR6 147 I WKUP
16 I/O D20 60 O CAS3 104 I/O ADR7 148 I RESET
17 VSS 61 VSS 105 VSS 149 VSS
18 I/O D19 62 O DOE 106 I/O ADR8 150 I IDLEEN
19 I/O D18 63 I CLKSEL2 * 107 I/O ADR9 151 I CLKSEL1
20 I/O D17 64 O ERROR 108 I/O ADR10 152 I CLKSEL0
21 I/O D16 65 O LOCK 109 I/O ADR11 153 I CLKEXT
22 I BTEST 66 I CTEST 110 I READY 154 I CLKIN
23 VDD 67 IO_VDD 111 VDD 155 IO_VDD
24 I/O D15 68 I BREQ 112 I/O ADR12 156 I IRQ11
25 I/O D14 69 O PBREQ 113 I/O ADR13 157 I IRQ10
26 I/O D13 70 O BGRNT 114 I/O ADR14 158 I IRQ9
27 I/O D12 71 I BMACK 115 I/O ADR15 159 I IRQ8
28 VSS 72 VSS 116 VSS 160 VSS
29 I/O D11 73 O BMREQ 117 I/O ADR16 161 N/A N.C.
30 I/O D10 74 O TIMER_OVF 118 I/O ADR17 162 N/A N.C.
31 I/O D9 75 O SAMEPAGE 119 I/O ADR18 163 N/A N.C.
32 I/O D8 76 I/O AS 120 I/O ADR19 164 N/A N.C.
33 IO_VDD 77 VDD 121 IO_VDD 165 N/A N.C.
34 I/O D7 78 I/O RDWR 122 I/O ADR20 166 N/A N.C.
35 I/O D6 79 O RDYOUT 123 I/O ADR21 167 N/A N.C.
36 I/O D5 80 O CS5 124 I/O ADR22 168 N/A N.C.
37 I/O D4 81 O CS4 125 I/O ADR23 169 N/A N.C.
38 I BMODE8 82 IO_VDD 126 I MEXC 170 IO_VDD
39 VSS 83 VSS 127 VSS 171 VSS
40 I/O D3 84 O CS3 128 I/O ADR24 172 N/A N.C.
41 I/O D2 85 O CS2 129 I/O ADR25 173 N/A N.C.
42 I/O D1 86 O CS1 130 I/O ADR26 174 N/A N.C.
43 I/O D0 87 O CS0 131 I/O ADR27 175 N/A N.C.
44 VDD 88 VSS 132 VDD 176 VSS
SPARClite Series 32-Bit RISC Embedded Processor
8
Fujitsu Microelectronics, Inc.
Table 3. Signal Descriptions
Symbol Type Description
CLKIN I
CLOCK:
The clock input pin. The clock is the timebase for the operation of the bus interface unit (BIU). An on-chip clock multiplier allows the CPU and core logic to run at integer multiples of the clock frequency (
×
1,
×
2,
×
3,
×
4, or
×
5).
CLKEXT I
EXTERNAL CLOCK BYPASS:
The external clock selection pin. If tied low, the clock is generated by the internal PLL/clock multiplier. If tied high, the external clock (i.e., the signal on CLKIN) is used without modification.
RESET I
SYSTEM RESET:
The reset input. The CPU and core logic are initialized by pulsing this input low. The clock must be stable for 100 ms before the reset pulse is de-asserted. The reset pulse must be a minimum of four CLKIN cycles in length. The CPU begins execution at location 0 three CLKIN cycles after the reset pulse is de-asserted.
CLKSEL0CLKSEL1CLKSEL2
I
INTERNAL CLOCK SELECT:
These pins select the clock frequency multiplier, as described in the table below.
Note: CLKSEL0 and CLKSEL1 do not have internal pullup resistors, so they must be tied high or low. CLKSEL2 has an internal 50K
Ω
pullup resistor.
CTESTBTEST
I
CTEST BTEST:
Test pins. Must be tied high.
ADR<27:2> I/O
ADDRESS BUS:
The 26-bit address bus ADR<27:2> references a 32-bit word. ADR1 and ADR0 are generated for 8- and 16-bit bus width trans-actions and are driven on the byte enables BE2 and BE3. The address is not valid during idle cycles. During bus grant mode, the address bus becomes an input, and it is used by the CS generator circuit and on-chip core logic. When ADR<27:2> is driven in this mode, ADR<31:28> are treated as 0 internally. If the DRAM controller is enabled, it multiplexes row and column addresses and drives them on ADR<13:2>.
ASI<3:0> I/O
ADDRESS SPACE IDENTIFIER:
The address space identifier (ASI) selects one of 16 separate address spaces referenced by the address bus. These spaces distinguish between user and supervisor space, instruction and data space, memory and peripheral control registers, and other addressable areas. The timing is identical to ADR<27:2>. The ASI signals become inputs during bus grant mode. This is used for CS generation and internal address decoding. In this mode, ASI<7:4> is treated as 0 internally by the CPU and core logic.
AS I/O
ADDRESS STROBE:
A one-cycle low pulse is driven on AS during the first clock of the bus cycle. The bus cycle starts with assertion of AS and ends with assertion of READY or RDYOUT. The AS signal is an input during bus grant mode and is used as an activation signal for CS generator circuits and wait-state generator circuits.
CS0CS1CS2CS3CS4CS5
O
CHIP SELECT:
The chip select signals are asserted if the address ranges indicated in the Address Range Specifier Register (ARSR) and the Address Mask Register (AMR) are referenced with the System Support Control Register (SSCR) CS enable bit (bit 4) set. An exception is CS0 (i.e. the boot ROM chip select), which has no Address Range Specifier Register and is always enabled. Each address range has a correspond-ing wait specifier which is used to automatically assert the READY signal after a user defined number of processor clocks. This allows a variety of memory and I/O devices with different access times to be connected to the processor without the need for additional logic.
D<31:0> I/O
DATA BUS:
This is the 32-bit data bus. It is a bidirectional data bus used for instruction fetch, data reads, and data writes. Instruction and word data must be aligned to 32-bit boundaries, and half words must be aligned to even addresses. Double words must be aligned to addresses which are multiples of 8. In 8-bit bus mode D<7:0> is used, and in 16-bit bus mode D<15:0> is used.
CLKSEL2 CLKSEL1 CLKSEL0 Internal Clock
H L L x1H L H x2H H L x3H H H x4L H H x5
MB86831
Fujitsu Microelectronics, Inc.
9
BE0
BE1
BE2/ADR1
BE3/ADR0
O
O
I/O
O
BYTE ENABLE:
Indicates the valid byte during a write when 32-bit bus width is used. During reads, BE<3:0> are asserted without regard to the bus width. When 8-bit bus width is used, BE2 and BE3 are driven with ADR1 and ADR0, respectively. When 16-bit bus width is used, BE2 is driven with ADR1. BE<3:0> are valid during the bus cycle period. During idle cycles, the output is undefined. During bus grant, the byte enables are high impedance, the DRAM controller may be enabled, and BE2 becomes the ADR1 input if 16-bit bus width is used.
*Notation such as (D<31:24>) shows the data bus bits being used.
RDWR I/O
READ/WRITE BUS TRANSACTION:
This signal is low during write cycles and high during read cycles and idle cycles. It is an input during bus grant mode, and it is used for generating DWE0 and DOE when the DRAM controller is enabled. This signal is not used in bus grant mode when the DRAM controller is disabled.
READY I
EXTERNAL READY:
This is a control signal asserted by the external memory system to indicate that the current bus transaction is being com-pleted and that it is ready to start with the next bus transaction in the following cycle. In case of a fetch from memory, the processor will strobe the value on the data bus at the rising edge of CLK_IN following the assertion of READY. For the case of a write, the memory system will assert READY when the appropriate access time has been met.In most cases, no additional logic is required to generate the READY signal. On-chip circuitry can be programmed to assert READY based on the address of the current transaction. The external system can override the internal ready generator to terminate the current bus cycle early. READY must be pulsed an appropriate number of times during a burst transfer or when multiple bus cycles are required to transfer data with 8- or 16-bit bus width.
Table 3. Signal Descriptions (Continued)
Symbol Type Description
Bus Width Access Type BE0,1,2,3
32-bit Write Byte 0 (D<31:24>) *1 0111Byte 1 (D<23:16>) 1011Byte 2 (D<15:8>) 1101Byte 3 (D<7:0>) 1110Half Word 0 (D<31:16>) 0011Half Word 1 (D<15:0>) 1100Word 0000
Read All data types 000016-bit Write Byte 0 (D<15:8>) 1000
Byte 1 (D<7:0>) 0100Byte 2 (D<15:8>) 1010Byte 3 (D<7:0>) 0110Half Word 0 (D<15:0>) 0000Half Word 1 (D<15:0>) 0010Word (D<15:0>) Access 0 0010Word (D<15:0>) Access 1 0000
Read Access 0 0000Access 1 0010
8-bit Write Byte 0 XX00Byte 1 XX01Byte 2 XX10Byte 3 XX11Half Word 0 Access 0 XX11Half Word 0 Access 1 XX10Half Word 1 Access 0 XX01Half Word 1 Access 1 XX00Word Access 0 XX11Word Access 1 XX10Word Access 2 XX01Word Access 3 XX00
Read Access 0 XX00Access 1 XX01Access 2 XX10Access 3 XX11
SPARClite Series 32-Bit RISC Embedded Processor
10
Fujitsu Microelectronics, Inc.
RDYOUT O
READY OUTPUT:
Assertions of the READY signal generated by any source, including the internal wait state generator, are visible to external devices on this pin. Internally generated READY assertions are synchronized to the clock. Externally generated READY assertions will appear on this signal with a small amount of propagation delay.
IRQ15/IRL3IRQ14/IRL2IRQ13/IRL1IRQ12/IRL0IRQ11IRQ10IRQ9IRQ8
I
INTERRUPT REQUEST/INTERRUPT REQUEST LEVEL:
These are interrupt inputs. Four of the inputs have a dual function. When the inter-rupt controller (IRC) is disabled (i.e. bits 0 and 1 in the IRC Mode Register are clear), IRL<3:0> is the encoded priority level of external interrupt requests, which compete with the on-chip interrupt sources for service by the CPU. Typically, IRL<3:0> would be generated by an external interrupt controller. Higher values have greater priority. IRL=0000
(2)
indicates no interrupt requests are pending, and IRL=1111
(2)
is defined by the SPARC architecture as a non-maskable interrupt. The external interrupt requests are sampled on two successive CLKIN clock periods to prevent false interrupts.When the interrupt controller is enabled, these pins are unencoded interrupt requests IRQ<15:8>. Each interrupt request can be programmed to be triggered on a high level, low level, rising edge, or falling edge. When an interrupt signal meets the qualifications to invoke an interrupt, an inter-rupt request is loaded in the IRC Request Sense Register. The IRC performs priority resolution and encoding to generate IRL<3:0>, which is passed to the CPU core.
BREQ I
BUS REQUEST:
When this signal is asserted by an external bus master, the CPU releases control of the bus after the current bus operation is completed. Certain operations require more than one bus cycle:(1) When an atomic load-store instruction is executed, the bus is released upon completion of both the load and the store.(2) In the case of a double word load or store, the bus is released if BREQ is asserted during the transfer of word 1 after the first word has been
transferred, and if BREQ is asserted during the transfer of word 2, the bus is released after transferring the second word.(3) Store in 8- and 16-bit bus widths:
The bus is released after transmission of the entire data object (for example, when a 32-bit word is transferred with 8-bit bus size, four 8-bit bus cycles occur before the bus is released).
(4) Load in 8 and 16-bit bus width:The bus is released after the entire word has been transferred.
BGRNT O
BUS GRANT:
This signal is asserted following a bus request (i.e. BREQ assertion) to indicate that control of the bus has been released to an external device.
PBREQ O
PROCESSOR BUS REQUEST:
This signal is asserted by the processor to indicate to an external bus arbiter that the CPU wants to regain control of the bus. This provides a handshake between the arbiter and the processor to allow the bus to be allocated based on demand.
LOCK O
BUS LOCK:
The CPU asserts this signal during execution of an atomic load-store. It indicates that the current bus transaction requires more than one bus cycle which cannot be split by releasing the bus to another bus master.
MEXC I
MEMORY EXCEPTION:
If this signal is low on the same clock that READY is asserted, the bus cycle is handled like a page fault, i.e. an instruc-tion or data access exception is invoked. This signal must not be asserted except on the same clock as READY. If a memory exception is signalled when the ET bit of the PSR is clear (i.e. traps are disabled), the CPU enters error mode.
ERROR O
ERROR SIGNAL:
This signal indicates that a trap has occurred while traps were disabled. When this happens, the CPU saves the PC and nPC to a register, loads the trap type in the TBR, and goes into error mode. Error mode can only be exited by a reset.
IDLEEN I
IDLE ENABLE:
When this pin is high and the previous cycle was loaded or stored to the CS0 area, the next cycle is started after insertion of a two-clock idle cycle. This is intended to accommodate EPROM boot memory with slow output disable, saving the addition of a buffer chip. When this pin is low, and a write cycle immediately follows a read cycle, an address cycle is inserted for one clock only (compatible with former versions of SPARClite).
BMODE8BMODE16
I
BOOT MODE8 and BOOT MODE16:
These signals select the bus width of the CS0 area. They are sampled during reset initialization. When BMODE8 is low, 8-bit bus width is selected, and when BMODE16 is low, 16-bit bus width is selected. When both pins are tied high, the bus width is 32 bits. The bus width of the CS1 through CS5 areas are not affected by these signals. The CS1 through CS5 areas are programmed through the Bus Width/Cacheable Control Register BWCR. Do not tie both of these inputs low.
NONCACHE I
NON-CACHEABLE:
This signal is asserted by external logic to indicate the data on the bus is non-cacheable. Low indicates non-cacheable and high indicates cacheable. This signal is used when the CBIR (Cache/BIU Control Register) Cacheability Enable bit (bit 7) is set. Normally, the NONCACHE signal is driven on a clock in which address strobe is asserted. However, if NONCACHE must be one clock or more late, it can be used by setting the Cache/BIU Control Register (CBIR) Noncacheable Wait State bits (bits 9,8). This signal is ignored during instruction fetches and when the internal cacheability is used.
PDOWN O
POWER DOWN:
This signal indicates transition to sleep mode (i.e. low power consumption mode) when it is low.
WKUP I
WAKE-UP:
This pin is driven low to break the CPU out of sleep mode. This pin is an asynchronous input, and must be driven with a pulse of 2 or more CLKIN periods. This pin must only be driven low when PDOWN is low, otherwise the behavior of the processor is undefined.
FLOAT I
FLOAT:
Driving this input low causes all output pins and bidirectional pins to go into high-impedance mode.
Table 3. Signal Descriptions (Continued)
Symbol Type Description
MB86831
Fujitsu Microelectronics, Inc.
11
Note: In the signal descriptions, names with an overbar indicate active low assertion. Dual function pins have two names separated by a slash (/).
BMREQ O
BURST MODE REQUEST:
This signal is asserted by the processor to indicate to the external system that the processor’s burst mode is enabled in the Bus Control Register (BCR) and the current transaction can be a burst. If the external system supports burst mode, BMACK can be asserted concurrently with READY to begin the burst mode transfer. BMREQ is asserted even when the DRAM burst enable bit in the System Support Control Register (SSCR) is set. However, in this case the internal DRAM controller drives the signal, so it is not necessary for external logic to drive BMREQ.
BMACK I
BURST MODE ACKNOWLEDGE:
This signal is asserted by external logic to indicate that it can support burst mode for the address currently on the bus. If driven low on or before the clock on which READY is asserted, burst mode data transfer is used. The signal can be driven low in the same clock as READY, or a clock before that and sustained until READY is driven. If the DRAM burst enable bit of the System Support Control Register (SSCR) is set, burst mode is used even if BMACK is not asserted.
SAMEPAGE O
SAME PAGE DETECT:
When bit 5 of the System Support Control Register (SSCR) is set, RAS page-hit detection is enabled. The address in the Same Page Master Register (SPSMR) and the previously accessed address are compared, and if they match SAMEPAGE is asserted. SAMEPAGE is never asserted in the first bus cycle following a transfer of bus control. The page size is specified in the Same Page Mask Register.
RAS0RAS1RAS2RAS3
O
DRAM ROW ADDRESS STROBE:
These are the row-address strobes from the DRAM controller.
CAS0CAS1CAS2CAS3
O
DRAM COLUMN ADDRESS STROBE:
These are the column address strobes from the DRAM controller. When 32-bit bus width is selected and 2-CAS DRAM configuration is used, CAS<0:3> correspond to byte 0 (b31-b24), byte 1 (b23-b16), byte 2 (b15-b8) and byte 3 (b7-b0). With 16-bit bus width and 2-CAS DRAM, CAS2 and CAS3 correspond to byte 0 (even byte address) and byte 1 (odd byte address), respectively. CAS0 and CAS1 are undefined when 16-bit bus width is selected. When 2-WE DRAM is used, CAS<0:3> are driven with identical signals.
DWE0DWE1DWE2DWE3
O
DRAM WRITE ENABLE:
These are the DRAM write enables. When a 2-WE DRAM configuration is used, DWE<0:3> correspond respectively to byte 0 (b31-b24), byte 1 (b23-b16), byte 2 (b15-b8) and byte 3 (b7-b0). When a 2-CAS DRAM is used, DWE<0:3> are driven with identical signals.
DOE O
DRAM OUTPUT ENABLE:
This is the output enable from the DRAM controller. DRAM interface is possible without using this signal for control of DWE0 and CAS<3:0> in early-write timing when a page-mode DRAM is used, but it is required for controlling the DRAM output drivers when EDO (hyper page-mode) is used.
TIMER_OVF O
TIMER OVERFLOW:
When the DRAM refresh timer is enabled in the System Support Control Register (SSCR), transparent DRAM refresh can be set up. The period between refresh cycles is programmed in the DRAM Refresh Timer Preload Register, and the signal pulse widths are pro-grammed in the DRAM Refresh Timer Register. TIMER_OVF is pulsed low when the timer counts down to zero. The timer is clocked by CLKIN, without frequency multiplication. Bit 31 of the DRAM Refresh Timer Preload Register controls the pulse width, either one clock or three clocks long. When a three-clock pulse width is selected, TIMER_OVF can be connected to an interrupt input (IRQx) of the interrupt controller.
Table 3. Signal Descriptions (Continued)
Symbol Type Description
SPARClite Series 32-Bit RISC Embedded Processor
12
Fujitsu Microelectronics, Inc.
Table 4. Pin Status Description Access Type
Pin During Reset During Bus Grant
ADR <27:2> O(X) I(D)
AS O(H) I(Z)
BE0 O(X) O(Z)
BE1 O(X) O(Z)
BE2/ADR1 O(X) O(Z), I(Z)
BE3/ADR0 O(X) O(Z)
CS<5:0> O(H) I(Z), O(Z)
ERROR O(H) O(V)
LOCK O(H) O(Z)
PDOWN O(H) O(H)
PBREQ O(H) O(V)
SAMEPAGE O(H) O(V)
D<31:0> I(Z) I(Z)
RDWR O(H) I(Z)
BGRNT O(H) O(L)
ASI<3:0> O(X) I(Z)
RDYOUT O(V) O(V)
BMREQ O(H) O(H)
TIMER_OVF O(H) O(V)
RAS<3:0> O(H) O(V)
CAS<3:0> O(H) O(V)
DOE O(H) O(V)
O(V) : Output driven to a valid level.
O(X) : Output is undefined.
O(Z) : Output is in high impedance.
O(H) : Output is driven high.
O(L) : Output is driven low.
I(Z) : Input is high impedance.
I(D) : If the DRAM controller is enabled, the address sampled on the assertion of AS will be driven from the next clock until READY is asserted. If the address maps to a DRAM area, a multiplexed address will be driven. If the DRAM controller is disabled, the signal will be in high impedance.
MB86831
Fujitsu Microelectronics, Inc.
13
Overview
The Fujitsu MB86831 is a high performance, 32-bit RISC processor with up to 66 or 80 MIPS peak performance at 66 or 80 MHz clock frequency, respectively. Like its predecessors, the MB86831 is based on the SPARC architecture and is upward code compatible with previous implementations. The MB86831 has been developed specifically for the needs of embedded applications that require high performance and high integration.
The MB86831 instruction set was designed for fast execution, with most instructions executing in a single cycle. The Integer Unit (IU) features a 5-stage pipeline which has been designed to handle data interlocks, an optimized branch handler for efficient control transfers, and a bus interface to handle single-cycle bus accesses to on-chip memory.
An internal register file consisting of 136 registers organized into eight overlapping windows provide rapid interrupt response time and context switches. The register file minimizes accesses to memory during procedure linkages and facilitates passing of parameters and assignment of variables.
On-chip 4-Kbyte, 2-way set-associative instruction and 2-Kbyte, 2-way set-associative data caches have been added to decouple the processor from external memory. These caches have been designed for maximum flexibility. For example, they allow cache lines to be locked for faster access to critical data.
Separate 32-bit on-chip instruction and data paths (i.e. Harvard-style architecture) provide a high-bandwidth interface between the IU and on-chip caches. These buses support single-cycle instruction execution as well as single cycle data transfers with the cache.
The MB86831 also includes hardware for integer multiply and divide step. The hardware support significantly improves the performance of these operations, with 32-bit integer multiplies executing in 5 clock cycles, 16-bit integer multiplies in 3 cycles, and 8-bit integer multiplies in 2 cycles.
Key Features
Fast Integer Unit Instruction Execution:
Simple operations make up the bulk of instructions in most programs, so execution speed can be greatly improved by designing these instructions to execute in as short a time as possible. In the SPARC architecture, the majority of instructions execute in one cycle with only a few of the more complex, such as integer multiply, taking additional cycles.
Large Register Set:
The large register set for the IU reduces the number of required accesses to data memory. The registers are organized in overlapping groups called register windows, which allow registers to be reserved for high priority tasks, such as interrupts, or for frequently called tasks such as operating system working registers. The overlapping windows also simplify parameter passing and reduce instruction overhead for procedure linkage.
On-Chip Caches:
Separate 2-way set associative instruction and data caches have been added to IU. This decouples the fast IU from off-chip memory because external memory access is only required on cache misses.
Cache Locking:
Both the instruction and data cache lines can be locked to ensure deterministic response and highest performance for critical or frequently called routines. Maximum flexibility has been designed into the cache to allow all or selected portions to be locked.
Bus Interface:
The requirement for glue logic between the MB86831 and the system is minimized by providing programmable chip selects, programmable wait state circuitry, programmable cacheable and non-cacheable memory address, and support for connection to fast page-mode DRAM or burst-mode EDO DRAM. Multiple bus masters are supported through a simple handshaking protocol. The MB86831 can boot from either 8-, 16-, or 32-bit wide memory. In addition, the programmable data bus allows reading/writing of different memory widths. For high frequency operation, the core can run at up to 5 times the bus. Note however, that the BIU frequency should not exceed 33MHz.
Clock Generator:
An external clock source must be supplied. Unlike some other members of the SPARClite family, there is no on-chip oscillator. A built-in phase-locked loop minimizes the skew between on and off-chip clocks.
Enhanced Instruction Set:
The MB86831 includes a fast integer multiply instruction which executes in 5, 3, or 2 cycles for 32-bit, 16-bit, or 8-bit multiplicands, respectively. An integer divide-step instruction cuts divide times by a factor of 10 over previous SPARC implementations. A scan instruction supports a single cycle search for the most significant 1 or 0 in a word or bit differing from sign bit.
SPARClite Series 32-Bit RISC Embedded Processor
14
Fujitsu Microelectronics, Inc.
Sleep Mode
The MB86831 has a Sleep Mode, i.e. a power-saving mode in which program execution is temporarily suspended. The PDOWN and WKUP pins and the Sleep Mode Register provide a mechanism to enter and exit Sleep Mode.
Table 5. MB86831 Instruction Set
CONDITION CODES UNCHANGED
ANDOR
XORANDNOT
OR NOTXNOR
CONDITION CODES UNCHANGED
ADDSUBTRACT
MULTIPLY(SIGNED/UNSIGNED)SCANSETHI
SHIFT LEFT LOGICALSHIFT RIGHT LOGICAL
SHIFT RIGHT ARITHMETIC
TO USER/SUPERVISOR SPACE SIGNEDLOAD BYTE
LOAD HALF-WORDLOAD WORD
LOAD DOUBLE WORDSTORE BYTE
STORE HALF-WORDSTORE WORD
STORE DOUBLE WORD
CONDITION CODES SETANDOR
XORAND NOTOR NOTXNOR
CONDITION CODES SETADD
SUBTRACTMULTIPLY(SIGNED/UNSIGNED)
MULTIPLY STEPDIVIDE STEP
TO USER SPACE UNSIGNEDLOAD BYTE
LOAD HALF-WORD
CONDITIONAL BRANCHCONDITIONAL TRAP
CALLRETURN
SAVERESTORE
JUMP AND LINK
EXTENDED AND CONDITION CODES UNCHANGEDADD
SUBTRACT
EXTENDED AND CONDITION CODES SETADD
SUBTRACT
TAGGED AND CONDITION CODES SET(WITH AND WITHOUT TRAP ON OVERFLOW)
ADDSUBTRACT
READ ASRWRITE ASR
READ PSRWRITE PSRREAD TBRWRITE TBR
TO ALTERNATE SPACE SIGNEDLOAD BYTE
LOAD HALF-WORDLOAD WORD
LOAD DOUBLE WORDSTORE BYTE
STORE HALF-WORDSTORE WORD
STORE DOUBLEWORD
TO ALTERNATE SPACE UNSIGNEDLOAD BYTE
LOAD HALF-WORD
ATOMIC OPERATION IN USER SPACESWAP WORD
LOAD/STORE UNSIGNED BYTE
ATOMIC OPERATION INALTERNATE SPACE
SWAP WORDLOAD/STORE UNSIGNED BYTE
READ WIMWRITE WIM
READ YWRITE Y
LOGICAL ARITHMETIC/SHIFT DATA MOVEMENT
CONTROL TRANSFER
READ/WRITE CONTROL REGISTER
MB86831
Fujitsu Microelectronics, Inc.
15
CPU
The MB86831 core is a high-performance, fully custom implementation of the SPARC architecture. The core is compact to leave chip real estate available for peripheral integration. The modular architecture allows the device family to be customized for varying application requirements. The core is made up of three functional units: the Instruction block, the Address block and the Execute block. (see Figure 1 below)
A five-stage instruction pipeline decodes the instructions and generates the control signals to the other blocks. The pipeline consists of Fetch (F), Decode (D), Execute (E), Memory (M), and Writeback (W) stages. Instruction memory is addressed and instructions are returned in the (F) stage, the register file is addressed and operands are returned in the (D) stage, the ALU produces results in the (E) stage, external memory is addressed in the (M) stage, and the register file is written back in the (W) stage.
Figure 1. MB86831 Integer Unit Data Path
Address Space
The MB86831 offers multiple large address spaces including separate user and supervisor spaces. In addition to 26 address lines, 8 alternate address space identifier bits (ASIs) distinguish between protected and unprotected space. Of the 256 possible ASI values, two select user instruction and data spaces, while the remaining ASI values define supervisor spaces. Included in the latter are peripheral control registers and direct access to the cache tag and data memories.
When a reset or trap occurs, the processor enters supervisor mode. In this mode, instructions and data come from supervisor space. While in supervisor mode, the processor has access to all protected ASI spaces. Four ASI spaces have been reserved for application-defined data spaces.
The distinction between user and supervisor space allows the hardware to protect against program errors. In developing real-time applications, for example, the separate spaces provide a mechanism for protecting the operating system from bugs in application code.
Registers
The MB86831 integer unit register set consists of both general-purpose registers and dedicated registers used for control and status.
The 136 general-purpose registers are divided into 8 global registers and 8 overlapping blocks or “windows”. Each window contains 24 registers. Of these, 8 are local to the window, 8 “out” registers overlap with the next window, and 8 “in” registers overlap with the previous window (see Figure 2 on page 16).
ir
e_ir
m_ir
w_ir
INSTRUCTION
adder
pc
0 TBR
inc (+4)
d_pc
e_pc
m_pc
ADDRESS
BA
ALU / SHIFTER
PSR/WIM/Y
st_align ld_align
read 1 read 2 read 3 writeREGISTER FILE
EXECUTE
I DATA
D ADDRESS D DATA
R Register
I ADDRESS
SPARClite Series 32-Bit RISC Embedded Processor
16
Fujitsu Microelectronics, Inc.
This organization makes it easy to pass parameters to subroutines. Parameters are written to the “out” registers and the subsequent procedure call decrements the window pointer to make a new set of registers available. The passed parameters are now available to the subroutine in the current window’s “in” registers.
Register windows improve performance in embedded applications because they function as local variable caches which retain either interrupt, subroutine, context, or operating system variables with no additional overhead. In addition, code space can be reduced by exploiting the efficient execution of procedure linkage.
The registers that make up the register file each have three read ports and one write port. The use of a four-port register file allows instructions to execute at one instruction per cycle, even in the case of the store instruction which can require reading up to three register operands.
The control and status registers include those defined by the SPARC architecture (see Table 11 on page 24) and those mapped into alternate address space to control peripheral functions (see Table 12 on page 25).
Figure 2. General-Purpose Register Window Organization
Instruction Sets
The MB86831 is upward-code compatible with other SPARC processors. Integer divide step, and scan for first changed bit have been added to the already powerful SPARC instruction set to improve performance in embedded applications. See Table 5 on page 14 for a list of the instructions.
Interrupt
A key measure of a processor’s suitability for use in an embedded application is its ability to handle interrupts with a minimum of latency and in a deterministic fashion. The MB86831 implemen-tation has been optimized to ensure not only low average latency but low maximum latency as well.
Interrupt response time is the sum of the time it takes the processor to finish its current task after recognizing an interrupt and the time it takes to begin executing interrupt service routine instructions. The MB86831 implements numerous features to minimize both factors.
To minimize the time it takes to finish the current task, the MB86831 is designed so that tasks can either be interrupted or completed in a minimum number of cycles. Implementation details that accomplish this include cache line misses that can be filled one word at a time through a prefetch buffer, integer divide that is interruptible through the use of a divide step instruction, fast multiply and a four-stage write buffer to defer pending bus transactions.
To minimize the time required to start executing the interrupt service routine, the processor switches to a new register window when an interrupt is detected. This feature allows the service routine to begin execution without first saving any registers on the stack. For even faster response, the application can also lock the service routine into the cache. This eliminates any latency caused by cache misses. The on-chip data cache can be used by the service routine as a fast local stack for minimum delay in accessing data.
Single vector trapping is a technique for saving code space and improving interrupt latency. When the SVT bit of the Ancillary State Register 17 is set, all traps vector through the first entry in the trap table rather than indexing to individual entries for each trap type. In some applications, this can allow the trap table to fit in the cache.
There are 15 different interrupt levels. The highest interrupt level is non-maskable.
w7ins
w7locals w7
outsw0outs w0
localsw6ins
w6locals
w6outs
w5ins
w5locals
w5outs
w4ins
w4localsw4
outsw3outs
w3locals
w3ins
w2ins
w2locals
w2outsw1
ins
w1locals
w1outs
CURRENT WINDOW
w0ins
globals
MB86831
Fujitsu Microelectronics, Inc.
17
Cache
The MB86831 has separate on-chip 2-way set-associative instruction and data caches. This allows the user to build a high-performance system without incurring the cost of requiring fast external memory and the associated control logic.
The caches use a buffered writethrough mechanism. Read hits can be satisfied by the caches without generating external bus cycles. Write hits are applied to both the cached data and the image of the data in external memory. Because writes to memory are buffered, the CPU can continue execution from cache without pausing to allow writes to external memory to complete. The caches are physically mapped.
The data cache is organized as two banks of 32 lines (see Figure 3 on page 18 for the organization of one bank). The instruction cache is organized as two banks of 64 lines (see Figure 4 on page 18 for the organization of one bank).
Lines are divided into sub-blocks each four bytes wide. On a cache miss, the caches are updated either 1 word (4 bytes) at a time, or 4 words at a time using the processor’s burst mode feature. Single-word updates minimize interrupt latency associated with long cache line replacements, while 4-word burst refills maximize the use of available bus bandwidth. An instruction pre-fetch buffer fetches the next sequential instruction anticipating that it will be needed to satisfy the next instruction cache miss.
The caches can be used in either normal mode or one of two lock modes. The two lock modes allow either the entire cache or just selected cache lines to be locked. The lock modes allow time-critical or performance-sensitive instructions and data to be locked in cache.
Global locking affects the entire content of either the instruction or data cache. Two control bits in the Cache/BIU Control Register enable or disable locking for either cache. With the entire cache locked, no valid cache line can be kicked out of the cache. To insure best performance however, invalid lines are allocated if possible. This is done automatically and incurs no time penalty.
Local cache locking makes it possible to dynamically lock selected instructions or data in the appropriate cache. This feature provides the flexibility, for example, to implement a known, deterministic response for certain critical interrupt routines by locking the routine’s code and data into the cache. Cache lines can also be locked to give priority to often used instructions or data which might otherwise be removed from cache.
In local lock mode, each entry can either be locked individually by software or automatically with hardware assist. For individual locking, software writes the lock bit in the appropriate cache tag line. For automatic locking, a bit in each Lock Control Register enables or disables the feature. The enable bit is set at the beginning of a routine for which the entries are to be locked. This causes the location of any cache access occurring while the bit is enabled to be locked into the cache. In addition to requiring just one initial cycle to enable, automatic entry locking incurs no overhead while in effect. Locked locations can be cleared with a single write to a control register.
In unlocked operation, the data cache uses a write-through update policy and allocates a cache line only on a load. Writes are buffered so that the processor can continue executing while data is written back to memory. In contrast, writes to locked data cache locations are not written through to main memory. Besides reducing external bus activity, this effectively configures a portion of data cache as on-chip RAM which does not map to external memory.
The data and instruction caches are designed to be accessed independently over separate data and instruction buses to allow data to be loaded from and stored to the cache at peak rates of 1 per instruction.
Different data memory spaces can be configured as cacheable or non-cacheable through either software programming or hardware control.
Following reset, bit 7 of Cache/BIU Control Register (ASI=0x01, ADR=0x0000 0000) is initialized so that cacheability is controlled by a hardware pin, NONCACHE. When the NONCACHE pin is low, the data associated with the address is non-cacheable, otherwise it is cacheable. In this mode, the hardware control of cacheability is independent of the chip selects.
The user can set bit 7 of Cache/BIU Control Register to allow software to control cacheability. By programming a few bits in the Bus Width and Cacheable Register (ASI = 0x01, address = 0x0000 016C), cacheability can be controlled by which chip select is used.
Cacheability for the CS4 and CS5 chip selects are special cases. When the internal DRAM controller is disabled, the cacheability of CS4 and CS5 is the same as the other chip selects. When the internal DRAM controller is enabled, the data memory space selected by CS4 is cacheable, and the data memory space referred by CS5 is non-cacheable. Unlike the other chip selects, CS5 can be programmed to
SPARClite Series 32-Bit RISC Embedded Processor
18
Fujitsu Microelectronics, Inc.
overlap the CS4 address range to define a noncachable region within DRAM.
Bus Interface
The bus interface unit (BIU) is designed for simplicity and high performance. Separate address and data buses make it easy to build fast systems. At the same time, on-chip circuitry allows these systems to be built with a minimum of external hardware.
The BIU runs at the rate of the external clock, however the CPU and core logic can run at rates of
×
1,
×2, ×3, ×4, or ×5 that rate. This is provided to ease the system design for applications where the CPU is running at a high frequency.
The bus interface supports fully programmable wait-state generation, address decoding with chip select outputs, booting from 8-, 16-, and 32-bit wide memory, and an auto-reload timer. A burst mode supports fast cache line fills. Address pins ADR<3:2> track the internal address changes during burst mode.
Each chip select can also be programmed to support 8-, 16-, or 32-bit wide memory. An exception is when CS4 and CS5 are used with the internal DRAM controller enabled, only 16- and 32-bit width is
supported for these areas. See the section on the DRAM controller for a more detailed description of DRAM access.
Interrupt ControllerThe interrupt controller (IRC) functions are a superset of the IRC functions of the MB86930 and MB86940 devices. It has four modes:
Trigger Mode Registers 0 and 1 set the trigger modes for each channel of channels 8 through 15 and channels 1 through 7, respectively. As shown in Table 6 below, an interrupt can be triggered by a high level, low level, rising edge, or falling edge.
Note: IRQx signals that do not have pins are tied low internally. If the trigger mode for these signals is set to low level, continuous interrupts will be generated.
Table 6. Interrupt Trigger Modes
TRGMD Trigger Mode
00 High Level (Initial Value)
01 Low Level
10 Rising Edge
11 Falling Edge
sub-block
7
sub-block
6
sub-block
4
sub-block
0block
010
31
Tag Entry Format
User/Supervisor lrulock
56 11
address tag
0123
31
sub-block
5
sub-block
3
sub-block
2
sub-block
1
1031Valid bits
BANK 1 or BANK 2
Figure 3. Data Cache Organization Figure 4. Instruction Cache Organization
sub-block
7
sub-block
6
sub-block
4
sub-block
0block
010
63
Tag Entry Format
User/Supervisor lrulock
56 11
address tag
0123
63
sub-block
5
sub-block
3
sub-block
2
sub-block
1
1031Valid bits
BANK 1 or BANK 2
MB86831
Fujitsu Microelectronics, Inc. 19
Request Sense Register (Read only)
31 16 0
Reserved (no bits) Request Sense 15-1
Bit 15-1 :REQSNS15-1 (initial value all 0)Shows channel 15-1 interrupt requests. If 1, indicates that an interrupt request is pending.
Bit 0 :Reserved (0 during reads)
15
IRC Mode Register
Table 7. Interrupt Controller Modes
31 16 2 1 0
Reserved (no bits) Reserved (reads as zero)
Bit1, 0 : IRCMD (Read/Write). (Initial value 00). IRC Mode Bit.
15
IRCMD
IRCMD IRC Mode
00 IRC disabled. IRL<3:0> inputs are selected.
01IRC enabled. IRL<3:0> are replaced by IRQ<15:12>, and the IRQ<11:8> inputs are available.
10 Reserved setting. Do not use.
11 Reserved setting. Do not use.
If events set by Trigger Mode Registers are detected, bits corresponding the events are set in the Request Sense Register. The Request Sense Register is read-only, and it is cleared by a reset. If an IRQx signal may be high following reset (i.e. the default trigger mode), its trigger mode should be programmed and any pending interrupt request cleared before enabling the interrupt controller.
Figure 5. Interrupt Controller Configuration
IRC
TriggerMode
ControlRequestSense
IRQMask
PriorityEncoder
IRLLatch
Selector
IRLMask
IRQ7IRQ6
::
IRQ1
IRLIN<3:0>/IRQ15-12
IRLOUT<3:0>
To
unused inputs tied low internally
CPU
IRQ11-8
SPARClite Series 32-Bit RISC Embedded Processor
20 Fujitsu Microelectronics, Inc.
The IRC Mode Register is used to enable the interrupt controller. When the IRCMD bits in this register are 00(2), the interrupt controller is disabled, and the encoded IRL<3:0> is passed to the IU as received from an external interrupt controller. When the IRCMD bits are 01(2), the interrupt controller is enabled, and the IRL<3:0> pins become interrupt request inputs IRQ<15:12>.
Interrupt Controller OperationThe interrupt controller is enabled by setting the IRCMD bits of the IRC Mode Register to 01(2). The interrupt requests then come from IRQ<15:1>. The number of interrupt inputs available at pins varies depending on how IRC resources are programmed. Interrupt requests are stored in the Request Sense Register when their trigger mode conditions are met. Of the stored interrupts not masked in the Interrupt Mask Register, those with the highest priority are encoded and stored in the IRL Latch/Clear Register.
If the IRL Mask (IM) bit in the Interrupt Mask Register is clear, the IRL is passed to the CPU. The interrupt is acknowledged by setting the CL bit of the IRL Latch/Clear Register, which clears the current IRL and allows the next interrupt level to be loaded.
If unmasked interrupt requests are pending in the Request Sense Register, they are loaded into the IRL Latch/Clear Register in order of priority.
If the IRCMD bits of the IRC Mode Register are 00(2), the interrupt controller is disabled and the signals on IRL<3:0> are passed to the CPU without modification.
1. Processing Interrupts as Traps
After reset, all of the mask bits in the Interrupt Mask Register are set (i.e. interrupts are masked). Software should then pro-gram the trigger modes and set bits 1 to 15 of the Request Clear Register to clear any pending interrupts. Next, the interrupt masks for the desired interrupts should be cleared and the IRL mask (bit 0 in the Interrupt Mask Register) should be cleared.
In the trap processing routine, software acknowledges the interrupt by first clearing the bit in the Request Clear Regis-ter corresponding to the interrupt, then clears the IRL latch by setting the CL bit in the IRL Latch/Clear Register. This allows the next pending interrupt request to be stored in the IRL latch.
2. Processing Interrupts By Polling
a. Polling the Request Sense Register Software can read the masked interrupt request sense bits in the Request Sense
Register and call the corresponding service routines when set bits are found. The service routine acknowl-edges the interrupt by clearing the bit through the Request Clear Register. This method is compatible with the method described above, with unmasked interrupts processed as traps and masked interrupts processed by polling.
b. Polling the IRL Latch/Clear RegisterThis is similar to processing interrupts as traps, except software polls the IRL Latch/Clear Register without generating traps. Unmasked interrupts are prioritized, encoded, and loaded into the IRL latch, however the IM bit in the Interrupt Mask Register is set, which prevents a trap from being called. Software can poll the IRL latch bits, call a service routine if it has any value other than 0000(2), then use the Request Clear Register to clear the corresponding bits in the Request Sense Register. The service routine acknowledges the interrupt by first clear-ing the bit in the Request Clear Register corresponding to the interrupt, then clears the IRL latch by setting the CL bit in the IRL Latch/Clear Register. This allows the next pending interrupt request to be stored in the IRL latch. Because the IM bit is set, this method is not com-patible with processing some interrupts as traps.
Clock GeneratorThe on-chip clock generator requires an external clock source (i.e. there is no on-chip oscillator). The external clock frequency is the same as the bus interface unit (BIU) operating frequency. The skew between the internal clock and an external input clock source is minimized through the use of an on-chip phase-locked loop.
The CPU and core logic can run at up to 5 times the frequency of external clock (Max. BIU frequency = 33MHz. This is enabled by the use of CLKSEL pins, as shown below.
Table 8. Clock Multiplication Factor
CLKSEL2 CLKSEL1 CLKSEL0 Internal Clock
H L L x1
H L H x2
H H L x3
H H H x4
L H H x5
MB86831
Fujitsu Microelectronics, Inc. 21
Figure 6. DRAM Controller Configuration
Programmable Chip SelectThe core logic includes six programmable chip selects with wait-state generator circuits. Chip select address ranges must not overlap. Some chip selects have assigned functions, as shown in Table 9, “Use of Chip Select,” on this page.
Chip select settings and wait state settings must be made while the cache is disabled. If the write buffer is operating with the cache enabled and these settings are changed, the bus interface unit may stop operating normally. ARSR (Address Range Specifier Register), AMR (Address Mask Register), SPGMR (Same Page Mask Register) and WSSR (Wait State Specifier Register) settings must be made during reset initialization, before enabling the cache.
Table 9. Use of Chip Select
ADR<25:1>
ASRDWR
BE<3:0>CS4
SAMEPAGETIMER_OVFDBURSTEN
OtherREADY
MPX
RAS<3:0>
CAS<3:0>
DWE<3:0>
DOE
RAS/CAS MPXTiming Control
ADR<13:2>
CPU CORE DRAMC
MB86831
CS0 Boot ROM area
CS1 General purpose
CS2 General purpose
CS3Control and status registers for external peripheral control and status registers
CS4 DRAM area
CS5Defines a noncacheable region when the noncache function is enabled and the built-in DRAM controller is enabled
DRAM Controller• High-speed page-mode DRAM support (burst mode and non-
burst mode supported)
• EDO (hyperpage) mode DRAM support (burst mode use only)
• Memory bus width: 16- and 32-bit bus width supported
• DRAM page size 256 to 4096 columns
• Self-refresh during sleep mode support
• Programmable RAS and CAS timing parameters
• DRAM controller enabled by setting the DRAM controller enable bit (bit 6) of the System Support Control Register (SSCR)
• Burst mode enabled by setting the DRAM burst enable bit (bit 7) of the System Support Control Register
SPARClite Series 32-Bit RISC Embedded Processor
22 Fujitsu Microelectronics, Inc.
CS0The CS0 output is a chip select for boot ROM. No Address Range Specifier Register exists for CS0. It has a base address of 0. Following reset, the CPU vectors to address 0. Unlike the other chip select outputs, the CS0 output defaults to an enabled condition after reset.
CS1 and CS2General-purpose chip selects with no special limitations.
CS3The CS3 area is assigned to peripheral control and status registers, and it is limited to a range of 1 Kbyte including both internal and external registers. When on-chip peripherals are accessed, the bus signals ADR<27:2>, AS, RDWR, BE<3:0>, and RDYOUT are driven with the same timing used for external bus cycles. Data also appears on D<31:0> when storing to on-chip peripherals selected with CS3.
The CS3 area can be set to 16- or 32-bit bus width, but 8-bit is not available. 32-bit bus-width peripherals can be assigned to CS3, and when using the Wait State Specifier Register, the Single Cycle Non Burst Mode bit and the Single Cycle Burst Mode bit must be clear. One or more wait states must be programmed, and the Override bit must be set. These rules apply to both on-chip and off-chip peripherals. On-chip peripherals, however, must be accessed with ADR<1:0> equal to zero using Half-Word Load and Half-Word Store instructions, because the on-chip peripherals only use D<15:0>.
CS4The CS4 area is assigned to DRAM. When the built-in DRAM controller is used, CS4 is the DRAM chip select. Even if an external DRAM controller is used, CS4 must be assigned to DRAM because the page-hit detection circuit (i.e. SAMEPAGE generator circuit) is enabled for the CS4 area. Whether the on-chip DRAM controller or an external controller is used, CS4 wait-state generation must be disabled. In the Wait State Specifier Register, the Wait Enable bit and the Single Cycle Non Burst Mode bit must be clear.
CS5When a noncacheable area is not present, CS5 can be used as a general-purpose chip select. When both a noncachable area and the on-chip DRAM controller are used, the CS5 area is forced to be noncacheable. Unlike the other chip selects, the CS5 address range may overlap the CS4 area, to define a noncacheable area in DRAM. These areas may overlap whether or not the DRAM controller is enabled. When CS5 is used in a DRAM area, the CS5 wait state generator must be disabled. In the Wait State Specifier Register, the Wait Enable bit and the Single Cycle Non Burst Mode bit must be clear. If the DRAM controller is enabled, the bus width of the CS5 area will be the same as that specified for the CS4 area.
Idle Cycle Insertion FunctionIDLEEN Pin Tied LowWhen the IDLEEN pin is tied low and a write cycle comes immediately after a read cycle, an idle cycle is automatically inserted for one CLKIN period.
IDLEEN Pin Tied HighWhen the IDLEEN pin is tied high and a bus cycle immediately follows an access to the CS0 area, an idle cycle is inserted for two CLKIN periods unless the second bus cycle also accesses the CS0 area. Because EPROM output disable time is long, this function can eliminate a tristate buffer in some designs. A single idle cycle is inserted when a DRAM write immediately follows a DRAM read, as shown in Table 10, “Idle Cycle Insertion Function,” on page 23.
MB86831
Fujitsu Microelectronics, Inc. 23
Table 10. Idle Cycle Insertion Function
Previous Cycle Following Cycle Inserted Idle Cycles
Instruction fetch from ROM Instruction fetch 0
Instruction fetch from ROM Data read 2
Instruction fetch from ROM Data read 2
Data read from ROM Instruction fetch 0
Data read from ROM Data read 2
Data read from ROM Data write 2
Instruction fetch from DRAM Instruction fetch 0
Instruction fetch from DRAM Data read 0
Instruction fetch from DRAM Data write 1
Data read from DRAM Instruction fetch 0
Data read from DRAM Data read 0
Data read from DRAM Data write 1
Data write from DRAM Instruction fetch 0
Data write from DRAM Data read 0
Data write from DRAM Data write 0
SPARClite Series 32-Bit RISC Embedded Processor
24 Fujitsu Microelectronics, Inc.
Table 11. MB86831 Control and Status Registers—Read/Write
n z v c reserved PILicc
Enable Trap (Enable=1, Disable=0, RST=0)
045678111219202324272831
0 0 0 010 0 1
n: (Negative=1, Non-Negative=0)z: (Zero=1, Non-Zero=0)v: (Overflow=1, No Overflow=0)c: (Carry=1, No Carry=0)
Processor Interrupt Level (Value 1-15, RST=Undefined)
Conditions
S MODE (Supervisor=1, User=0, RST=Undefined)
Prior S Mode
Current Window Pointer (Value=0-7, RST=Undefined)
WIMWindow Invalid Mask
0567831
reserved
Window Invalid Mask (Invalid=1, Valid=0, RST=Undefined)
4 3 2 1
TBRTrap Base Register
0111231Trap Base Address
4 3Trap Type
NULL
YY Register
031
ASR 17
031
reserved (must be zero on writes)
1
Single Vector Trapping (Enabled=1, Disabled=0, RST=0)
(RST=Undefined) (RST=0)
Reserved
3 2
PCProgram Counter
031
NPCNext Program Counter
031
used for MULScc, DIVScc, MU
instruction address
instruction address
w0w5w6w7 w4 w3 w2 w1
Process State Register PSR
Ancillary State Register 17
MB86831
Fujitsu Microelectronics, Inc. 25
Table 12. MB86831 Memory Mapped Control Registers—Read/Write
ASI ADDRESS
0x 1 0x 0000 0000
Cache/BIU Control 031
Instruction Cache Enable (Enabled=1, Disabled=0, RST=0)
Write Buffer Enable (Enabled=1, Disabled=0, RST=0)Prefetch Buffer Enable (Enabled=1, Disabled=0, RST=0)
Global Data Cache Lock (Lock=1, Unlock=0, RST=0)Data Cache Enable (Enabled=1, Disabled=0, RST=0)
Global Instruction Cache Lock (Lock=1, Unlock=0, RST=0)
ASI ADDRESS
0x 1 0x 0000 0004
Lock Control
Data Cache Entry Auto Lock (Disabled=0, Enabled=1, RST=0)Instruction Cache Entry Auto Lock (Disabled=0, Enabled=1, RST=0)
31
0x 1 0x 0000 0010Restore Lock Control Register (Restore=1, Ignore=0, RST=0)
31
0x 1 0x 0000 0020Data Burst Enable (Enable=1, Disable=0, RST=0)
Instruction Burst Enable (Enable=1, Disable=0, RST=0)
31
31
reserved
Same Page Enable (Enabled=1, Disabled=0, RST=0)Chip Select Enable (Enabled=1, Disabled=0, RST=0)
Programmable Wait-State (Enabled=1, Disabled=0, RST=1)Timer On/Off (Enabled=1, Disabled=0, RST=0)
123456
Reserved
01
0
01
0123456
DMA Cycle Steal (Enabled=1, Disabled=0, RST=0)Parity (Odd=1, Even=0, RST=0)
ASI ADDRESS
0x 1 0x 0000 0008
Lock Control Save
Previous Instruction Cache Auto Lock (Off=0, On=1, RST=0)
31 0
0x 1 0x 0000 000CAuto Lock Failed (False=0, True=1, RST=0)
31 0
1
Previous Data Cache Auto Lock (Off=0, On=1, RST=0)
reserved
reserved
reserved
reserved
reserved
reserved
7
Cacheability Enable (Enabled=1, Disabled=0, RST=0)
DRAM controller enable
7
89
Non-Cacheable Wait-state
DRAM Burst enableASI ADDRESS
0x 1 0x 0000 0080
ASI ADDRESS
Cache Status
ASI ADDRESS
Restore Lock Control
ASI ADDRESS
Bus Control
System Support ControlRegister (SSCR)
SPARClite Series 32-Bit RISC Embedded Processor
26 Fujitsu Microelectronics, Inc.
Table 12. MB86831 Memory Mapped Control Registers—Read/Write (continued)
1. This register is write only
ASI ADDRESS
0x 1 0x 0000 01600x 0000 01640x 0000 0168
Single Cycle Burst Mode(On=1, Off=0, RST=0)
Single Cycle Non Burst Mode(On=1, Off=0, RST=0)
Wait Enable (On=1, Off=0, RST=0)
ASI ADDRESS
0x 1
CS1,CS0CS3,CS2CS5,CS4
0x 0000 01400x 0000 01440x 0000 01480x 0000 014C0x 0000 01500x 0000 0154
CS0CS1CS2CS3CS4CS5
Override (On=1, Off=0, except CS0, RST=1)
0131 30 23 22
ADR <31:10> Mask(0=Care, 1=Don’t Care, RST=Undefined)
031 27 26 25 24 23 22 21 2019 18 14 13 9 7 6 5 4 3
(RST=Undefined)
Count1(RST=
Undefined)
Count2(RST=
Undefined)
Count1(RST=
Undefined)
Count212
ASI ADDRESS0x 1 0x 0000 0124
0x 0000 01280x 0000 012C0x 0000 01300x 0000 0134
Address RangeSpecifier Register (ARSR)1
NOTE: CS0 is hardwired to ASI=0x9 ADR<31:10> = <0..0>
CS1CS2CS3CS4CS5
0131 30 23 22
ASI<7:0> ADR<31:10>(RST=Undefined) (RST=Undefined)
ASI ADDRESS
0x 1 0x 0000 0120
0131 30 23 22
[Care=0, Don’t Care=1, RST=0] [Care=0, Don’t Care=1, RST=0]
NOTE: CS0 ADR<14:10> = 1, ADR<31:15> = 0, ASI = 0x9 at reset.
8
Reserved
Wait State Specifier (WSSR)
Address Mask (AMR)
ASI<7:0> Mask
Same Page Mask (SPGMR)
ASI<7:0> Mask ADR<31:10> Mask
MB86831
Fujitsu Microelectronics, Inc. 27
Table 12. MB86831 Memory Mapped Control Registers—Read/Write (continued)12131415161718192021222324
ASI ADDRESS
0x 1 0x 0000 016C
Bus Width and Cacheable 31
reserved
Cacheable (0= cacheable, 1=noncacheable)Internal /External cacheable (0=NONCACHE_ pin, 1=internal)
CS5 CS4 CS3 CS2 CS1 CS0
01234567891011
CS5 CS4 CS3 CS2 CS1
RSVD
Bus Width Control (bit3 = BW0, bit2 = BW1)
ASI ADDRESS
0x 1 0x 0000 0178
DRAM Refresh Timer Pre-Load 031 16 15
reserved Timer Pre-Load Value(RST=Undefined)
Ancillary Version Register (V E R 2) [Read Only]
31 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
bit31-16 : Reserved [Read undefined]bit15-0 : MB86831 Version Number (Value=0)
16 15
Reserved
Sleep Mode Register (S L P M D) [Write Only]
31 0
bit31-1 : Reserved bit0 : Sleep Mode (On=1, Off=0, RST=0)
1
Reserved
30
bit31 : 3Cycle Mode (On-1, Off=0, RST=0)bit30-16 : Reserved [“0” write, read undefined]bit15-0 : Timer Pre-load Value (RST=0xffff)
0x 1 0x 0000 0174
031 16 15
reserved Timer Value(RST=Undefined)
bit31-16 : Reserved [“0” write, read undefined]bit15-0 : Timer Value (RST=undefined)
ASI ADDRESS
DRAM Refresh Timer
ASI=0x01, Address=0x00020000
ASI=0x01, Address=0x00020004
SPARClite Series 32-Bit RISC Embedded Processor
28 Fujitsu Microelectronics, Inc.
Table 12. MB86831 Memory Mapped Control Registers—Read/Write (continued)
Request Clear Register (R E Q C L R) [Write Only]
IRC Mode Register (I M O D E) CS3=L, Address<9:2>=0x06
Trigger Mode1 Register (T R G M1 )
bit31-16 : Reserved [“0” write, read undefined]bit15-2 : Trigger Mode (High Level=00, Low Level=01, High Edge=10, Low Edge=11, RST=00)bit1-0 : Reserved [“0” write, read “0”]
31 8 7 6 5 4 3 2 1 0
Reserved
10 916 15 14 13 12 11
ch7 ch4 ch3 ch2 ch1 00ch5ch6
Request Sense Register (R E Q S N S) [Read Only]
bit31-16 : Reserved [“0” write, read undefined]bit15-1 : Request Sense 15-1 (RST=0)bit0 : Reserved [“0” write, read “0”]
31 1 0
Reserved
CS3=L, Address<9:2>=0x02
16 15Request Sense 15-1
IRL Latch/Clear Register (I R L A T)
bit31-16 : Reserved [“0” write, read undefined]bit15-5 : Reserved [“0” write, read “0”]bit4 : IRL Clear [write only] Clear=1, Not Clear=0bit3-0 : IRL Latch [read only] (RST=0000)
31 3 0
Reserved
CS3=L, Address<9:2>=0x05
16 15Reserved IRL
5 4
CL
bit31-16 : Reserved [“0” write, read undefined]bit15-2 : Reserved [“0” write, read “0”]bit1-0 : IRC Mode [I R C M D] (Disable=00, Enable=01, RST=00)
31 1 0
Reserved16 15
Reserved IRCMD
2
Interrupt Mask Register (I M A S K)
bit31-16 : Reserved [“0” write, read undefined]bit15-1 : Mask 15-1 (Mask=1, Not Mask=0, RST=1)bit0 : IRL Mask (Mask=1, Not Mask=0, RST=1)
31 1 0
Reserved
CS3=L, Address<9:2>=0x04
16 15Mask 15-1 IM
bit31-16 : Reserved [“0” write]bit15-1 : Request Clear 15-1 (Clear=1, Not Clear=0)bit0 : Reserved [“0” write]
31 1 0
Reserved16 15
Request Clear 15-1
Trigger Mode0 Register (T R G M 0)
bit31-16 : Reserved [“0” write, read undefined]bit15-0 : Trigger Mode (High Level=00, Low Level=01, High Edge=10, Low Edge=11, RST=00)
31 8 7 6 5 4 3 2 1 0Reserved
10 9
CS3=L, Address<9:2>=0x00
16 15 14 13 12 11
ch15 ch12 ch11 ch10 ch9 ch8ch13ch14
CS3=L, Address<9:2>=0x01
CS3=L, Address<9:2>=0x03
MB86831
Fujitsu Microelectronics, Inc. 29
Table 12. MB86831 Memory Mapped Control Registers—Read/Write (continued)DRAM Bank Configuration Register (D B A N K R)
bit31 : Access Error (ERR) (Error=1, No Error=0, RST=0, “0” write Clear)bit30-11 : Reserved [“0” write, read undefined]bit10-9 : DRAM Start Address [STADR] (RST=0)bit8 : Hyper Page Enable [HE] (Page Mode DRAM=0, EDO DRAM=1, RST=0)bit7 : DRAM Type [TP] (4CAS-1WE=0, 4WE-1CAS=1, RST=0)bit6-4 : Column Address [COL] (RST=011)bit3-0 : Bank Size [BKSIZE] (RST=0011)
31 8 7 6 4 3 0
Reserved10 9
CS3=L, Address<9:2>=0x08
11ERR TP COL BKSIZEHESTADR
DRAM Timing Register (D T I M R)
bit31-5 : Reserved [“0” write, read undefined]bit4 : R-AS Pre-Charge Time specifier bit during Self-Refresh [TRPS] (2Cycle=0, 4Cycle=1, RST=1)bit3-2 : R-AS Pulse Width specifier bit during CBR Refresh [TRASCBR] (1cycle=00, 2cycle=01, 3cycle=10, RST=01)bit1 : C-AS Pulse Width specifier bit [TCAS] (1Cycle=0, 2Cycle=1, RST=1)bit0 : R-AS Pre-Charge Width specifier bit [TRP] (1Cycle=0, 2Cycle=1, RST=0)
31 5 4 3 2 1 0
Reserved TRASCBR TCAS TRPTRPS
CS3=L, Address<9:2>=0x09
SPARClite Series 32-Bit RISC Embedded Processor
30 Fujitsu Microelectronics, Inc.
Table 12. MB86831 Memory Mapped Control Registers—Read/Write (continued)
ASI ADDRESS
0x C
Instruction Cache Tag
Sub Block Valid (Valid=1, Invalid=0, RST=Undefined)User/Supervisor (User=0, Supervisor=1, RST=Undefined)
Least Recently Used (RST=Undefined)Entry Lock (Locked=1, Unlocked=0, RST=Undefined)
ASI ADDRESS
0x E
Data Cache Tag
01245611 1031
ADDRESS TAG [RST=Undefined]
Sub Block Valid (Valid=1, Invalid=0, RST=Undefined)User/Supervisor (User=0, Supervisor=1, RST=Undefined)
Least Recently Used (RST=Undefined)Entry Lock (Locked=1, Unlocked=0, RST=Undefined)
01245691031
ADDRESS TAG [RST=Undefined]
0x 0000 0000
0x 0000 07E0
Bank 1
0x 8000 0000
0x 8000 07E0
Bank 2
0x 0000 0000
0x 0000 03F0
Bank 1
0x 8000 0000
0x 8000 03F0
Bank 2
31
Cache LRU, Lock Bit Clear (Write Only) (Clear = 1, Not Clear = 0)Valid Bit Clear (Write Only) (Clear = 1, Not Clear = 0)
ASI ADDRESS
0x C
Instruction Cache Invalidate 012
31
Cache LRU, Lock Bit Clear (Write Only)Valid Bit Clear (Write Only)
ASI ADDRESS
0x E
Data Cache Invalidate 012
0x 0000 10000x 8000 1000
Bank 1Bank 2
0x 00001 000 0x 80001 000
Bank 1Bank 2
by 8
by 8
by 8
by 8
ASI ADDRESS
0x 3
Data Tag Lock Bits
Entry Lock (Locked=1, Unlocked=0, RST=Undefined)0x 0000 0000
0x 0000 03F0
Bank 1
0x 8000 0000
0x 8000 03F0
Bank 2
031
by 8
by 8
reserved
reserved
reserved
reserved
ASI ADDRESS
0x 2
InstructionTag Lock Bits
Entry Lock (Locked=1, Unlocked=0, RST=Undefined)0x 0000 0000
0x 0000 07E0
Bank 1
0x 8000 0000
0x 8000 07E0
Bank 2
031
by 8
by 8
reserved
ASI ADDRESS0x D
Instruction Cache Data
0x 0000 0000
0x 0000 07E0
Bank 1
0x 8000 0000
0x 8000 07E0
Bank 2
031
by 8
by 8
Address Tag (RST=undefined)
ASI ADDRESS0x F
Data Cache Data
0x 0000 0000
0x 0000 03F0
Bank 1
0x 8000 0000
0x 8000 03F0
Bank 2
031
by 8
by 8
Address Tag (RST=undefined)
MB86831
Fujitsu Microelectronics, Inc. 31
Bus OperationThe bus interface unit (BIU) is the interface to external devices. It includes the address and data buses, the interrupt request bus, and various control signals. At any time, the BIU is either handling transfers to or from off-chip devices, arbitrating for bus access, or idle.
Operation of the BIUOn writes to external memory, the BIU uses a four-word write buffer. When the BIU receives a request for a write transaction it stores the write data and address in the write buffer, which releases the IU to continue executing out of on-chip cache. The BIU then proceeds to complete the write to external memory. In most cases, the write buffer will hide external memory latency from the IU. The exceptions are in cases where the write buffer is full from previous transactions or if the subsequent IU cycle results in an instruction cache or data cache read miss. In these cases, IU execution is held until the write buffer is emptied.
The write buffer operates only when both the instruction and data caches are enabled. When the bus is granted to an external bus master, a store to the write buffer does not cause the assertion of PBREQ. This allows the external bus master to continue operating while the CPU is executing out of the on-chip caches.
The BIU includes a one-stage prefetch buffer for instruction fetches. This buffer is used to fetch the next sequential instruction after an instruction cache miss. The instruction is prefetched only if the BIU does not have a request for a bus transaction from the IU and no external device is requesting use of the bus. The prefetch buffer operation is suspended if the buffer is full. This occurs if the prefetched instruction is a hit in the instruction cache. The buffer restarts after another instruction cache miss. If an exception occurs during an instruction prefetch, the exception is not sent to the IU unless the instruction is actually requested by the IU. The prefetch buffer operates only when the instruction cache is enabled.
In any cycle, the BIU can receive a request for access to either or both instruction and data memory. If it receives a request for both in the same cycle, it completes the data memory transaction first.
Exception HandlingThe external memory system can indicate an error during a memory operation. The BIU signals the appropriate data or instruction exception to the IU which calls the appropriate trap.
As mentioned above, the IU can continue operation after putting the data and address for a store in the write buffer. If an exception is detected while completing this buffered write, then the BIU signals a data access exception to the IU.
Any system which needs to recover from this error should store the address and data of these write transactions in hardware. If the system can generate both read and write exceptions, then the system must also provide a status bit which indicates whether the exception was generated on a read or a write transaction. With this information, the data access exception service routine can determine the cause of the exception and recover.
If the write buffer is operating, an exception can potentially cause other exceptions due to the flush of the four write buffer levels. A system that needs the ability to recover must store up to four separate sets of address and data.
Bus CyclesFigure 7 on page 34 through Figure 28 on page 45 illustrate representative combinations of bus cycles.
LoadRegardless of the external bus width (8, 16, or 32 bits), all instruction fetches and data reads (including load byte and load half word) load a 32-bit quantity. This is done for compatibility with MB8693x processors with data cache where the smallest granularity in the cache is one word. Bus width can be programmed based on chip select areas to be 8, 16, or 32 bits.
Load (32-bit bus width)Whenever a load from data memory is requested or an instruction cache miss occurs, the BIU performs a read from external memory (see Figure 7 on page 34).
With a 32-bit external data bus, a read transaction begins with the BIU asserting AS, to indicate a new bus transaction. The AS signal is de-asserted after one cycle. At the same time the ADR<27:2> and ASI<3:0> bits are driven with the location to be read. The BIU drives the RDWR signal high to indicate a read transaction. Because all loads transfer 32 bits, BE<3:0> are all driven low.
The external memory system responds with the read data on pins D<31:0>. It also asserts the READY signal when the external device is ready for the bus cycle to complete. For slow memory, the READY signal can be delayed until data is valid.
A load double operation is treated as two back-to-back word reads.
SPARClite Series 32-Bit RISC Embedded Processor
32 Fujitsu Microelectronics, Inc.
Load (16-bit bus width)When the bus is programmed to be 16 bits wide (defined by the chip select region) every load will transfer 32 bits. Figure 17 on page 40 shows a load (byte, half word, word) operating with an 16-bit bus. For the load byte and load half word instructions, the IU masks off the bits which are not required. For a 16-bit bus, the BE2 pin is defined to be the ADR1 address bit. BE<1:0> and BE3 are unused and driven low.
Load (8-bit wide bus)When the bus is programmed to be 8 bits wide (defined by the chip select region) every load will transfer 32 bits. Figure 15 on page 39 shows a load (byte, half word, word) operating with an 8-bit bus. For the load byte and load half word instructions, the IU masks off the bits which are not required. For an 8-bit bus BE<3:2> are the ADR<0:1> address bits, respectively. BE<1:0> are unused and are driven to undefined states.
Load with ExceptionIf the external memory system sees a memory exception it can terminate the current memory transaction by asserting the MEXC and READY signals. The data on the data bus is ignored. See Figure 8 on page 35.
StoreUnlike loads, which always transfer 32 bits, only the minimum number of bus cycles required to complete the store are performed. For example, only two bus cycles are required to do a half-word store on an 8-bit bus.
Store (32-bit bus width)A write transaction begins with the BIU asserting AS, to indicate the start of a new bus transaction. The AS signal is de-asserted after one clock. At the same time, the ADR<27:2> and ASI<3:0> pins are driven with the location to be written while the D<31:0> pins carry the write data. BE<3:0> indicate which bytes to write for a given type of store operation (byte, half-word, or word). The BIU drives the RDWR signal low to indicate a write transaction. See Figure 9 on page 35.
The external memory system responds with the assertion of READY when it has stored the data. Or, if the internal wait-state generator is enabled, READY is generated internally.
A store double operation is treated as two back-to-back word writes.
Store (16-bit wide bus)Stores to 16-bit memory are sized to the bus. That is, for a 16-bit bus, a store word requires two cycles while a store halfword or store byte requires a single cycle. Figure 18 on page 40 shows the timing for different types of stores. For a 16-bit bus, BE2 is driven with ADR<1>. BE3 is unused and is driven low. BE<1:0> are defined to be the high and low order byte enables, respectively.
Store (8-bit wide bus)Stores to 8-bit memory are sized to the bus. That is, for a 8-bit bus, a store word requires four cycles, a store halfword requires two cycles, and store byte requires a single cycle. Figure 17 on page 40 shows the timing for different types of stores. For an 8-bit bus, BE<2:3> are driven with ADR<1:0>. BE<1:0> are unused and are driven to undefined states.
Store with ExceptionIf an access exception occurs on a write, the external memory system can terminate the current memory transaction by asserting the MEXC and READY signals. The external memory system is expected to ignore the data on the data bus in this situation. See Figure 10 on page 36.
Atomic Load StoreAn atomic load store executes as a load followed by a store with no operation allowed in between. The LOCK signal is asserted to indicate that the bus is being used for more than one external memory operation. See Figure 11 on page 36.
There is one idle cycle between the termination of the read and the beginning of the write to provide time for changing the direction of the data bus drivers.
External Bus Request and GrantAny external device can request ownership of the bus by asserting the BREQ signal. When control of the bus is granted, the BIU asserts the BGRNT signal and floats its bus drivers. In the following cycle, the external device can begin its transaction. On completion of its transaction the external device de-asserts the BREQ signal. The BIU responds by de-asserting the BGRNT signal in the following cycle. See Figure 12 on page 37.
A separate signal, PBREQ, is asserted by the processor to indicate to a bus arbiter that it has a pending bus transaction. This allows the bus to be allocated based on demand. The signal, PBREQ, is asserted when the write buffer is full or the CPU is doing an instruction or data fetch. The CPU is the default owner of the bus.
MB86831
Fujitsu Microelectronics, Inc. 33
8- and 16-Bit Bus ModesAny chip select (CS<5:0>) can be mapped to 8-, 16-, or 32-bit bus width, however CS4 and CS5 cannot be mapped to 8-bit bus width if the DRAM controller is enabled. Memory width for CS0 is selected at system reset with the BMODE8 and BMODE16 signals. Table 13, below shows the bus width options available for the CS0 area. Memory width for the CS<5:1> areas is programmed by two bits for each chip select in the Bus Width and Cacheable Register. Table 14, also below, shows the programming bits and the corresponding bus width of each chip select.
Table 13. Bus Width Control of CS0
Table 14. Bus Width Control Bits of CS1 to CS5
8- and 16-bit transactions are similar to 32-bit transactions except that AS is asserted only once at the beginning of the bus cycle for a load operation, and READY is asserted at the end of each byte or halfword transfer. The BE<3:0> signals indicate the byte or halfword being read or written (see Figure 15 on page 39 through Figure 18 on page 40).
For 32-bit writes to 8- or 16-bit memory and 16-bit writes to 8-bit memory, the BIU drives BE<2:3> as ADR<1:0>, and initiates multiple transactions.
When the internal DRAM controller is enabled, 8-, 16-, and 32-bit bus width are available for CS0 through CS3, and only 16- or 32-bit memory bus are available for CS4 and CS5, depending on the DRAM bus width.
Burst Mode TransactionsFor systems that can support burst mode transactions, the DRAM controller can be programmed to support four-word bursts. When burst mode is enabled, BMREQ is asserted at the beginning of each bus cycle for which a burst transfer is allowed (see Table 15, below). If the memory system can support a burst for the current bus address, the memory system asserts BMACK to begin the burst transaction. BMACK is asserted on the first word of the burst transaction only. READY is asserted for each word of the burst. Systems that do not support burst mode for the current address should not assert BMACK. If either BMREQ or BMACK is not asserted for a transaction, only one word is transferred for each assertion of AS.
Table 15. ADR<3:2> Sequence in Burst Mode
Selection of Hyperpage Mode (EDO Mode)By setting bit 8 of the DRAM Bank Configuration Register, hyperpage DRAM can be connected and controlled. However, if the DRAM Burst Enable Bit (bit 7 of the System Support Control Register) is clear, access timing is controlled in the same cycle as normal page-mode DRAM.
When the DRAM Burst Enable Bit is set and hyperpage is enabled, hyperpage burst access is performed. In this case, ADR<3:2> is not a value from the CPU but is generated by the DRAM controller itself. Because data is driven before it is received from the DRAM, the CAS cycle time can be shortened compared to the normal page mode burst transactions.
BMODE16 BMODE8 Bus Width
0 0 Illegal
0 1 16-bit Bus
1 0 8-bit Bus
1 1 32-bit Bus
BW1 BW0 Bus Width
0 0 32-bit Bus
0 1 8-bit Bus
1 0 16-bit Bus
1 1 Illegal
Bus Cycle 1 Bus Cycle 2 Bus Cycle 3 Bus Cycle 4
00 01 10 11
01 00 11 10
10 11 00 01
11 10 01 00
SPARClite Series 32-Bit RISC Embedded Processor
34 Fujitsu Microelectronics, Inc.
Basic DRAM Access Timing
Table 16. DRAM Access Timing
Figure 7. Typical Back-to-Back Loads (Same as Load Double)
Bus Width Burst Mode Page/EDO DRAM Mode R/W Timing Chart
32-bit
Off PageRead See Figure 19
Write See Figure 19
On
PageRead See Figure 20
Write See Figure 20
EDORead See Figures 21 and 22
Write See Figure 22
16-bit
Off PageRead See Figure 23
Write See Figure 23
On
PageRead See Figure 24
Write See Figure 27 (Page and EDO have same timing)
EDORead See Figures 25 and 26
Write See Figure 27 (Page and EDO have same timing)
CLKIN
ADR<27:2>ASI<3:0>
AS
RDWR
READY
LOAD 1 LOAD 2
A2A1
D<31:0> D1 D2
MB86831
Fujitsu Microelectronics, Inc. 35
Figure 8. Load with Exception
Figure 9. Typical Back-to-Back Stores (Same as Store Double)
CLKIN
ADR<27:2>ASI<3:0>
AS
RDWR
READY
LOAD 1
A2A1
D<31:0> INVALID
MEXC
CLKIN
ADR<27:2>ASI<3:0>
AS
RDWR
READY
STORE 1 STORE 2
A2A1
D<31:0> D1 D2
SPARClite Series 32-Bit RISC Embedded Processor
36 Fujitsu Microelectronics, Inc.
Figure 10. Store with Exception
Figure 11. Atomic Operation
CLKIN
ADR<27:2>ASI<3:0>
AS
RDWR
READY
STORE 1
A2A1
D<31:0>
MEXC
Note: A load followed by a store requires an intervening clock cycle on the bus while a store followed by a load can occur in consecutive clock cycles.
ADR<27:2>ASI<3:0>
LOAD 1 STORE 1
A1 A2 A3
CLKIN
AS
RDWR
READY
LOCK
D2D1D<31:0>
MB86831
Fujitsu Microelectronics, Inc. 37
Figure 12. Bus Request and Grant Cycle
Figure 13. Timing During a DRAM Read During Bus Grant Mode
CLKIN
BGRNT
Processor Bus Cycle n Complete
BREQ
Processor Bus Cycle n+1 Start
All Bus Drivers High-Impedance
CLKIN
ASI
D<31:0>
RDWR
BGRNT
AS
BREQ
Page MISS-HITPage Hit
Idle
VALID
Row Column Column
Pullup required, as BE2 is input mode
ADR<27:2>ADR1/(BE2)
CS4
RDYOUT
RASx
CASx
BE<3:0>
CPU Driving
External Circuit driving
Floating
DRAM output DRAM output
VALID
SPARClite Series 32-Bit RISC Embedded Processor
38 Fujitsu Microelectronics, Inc.
Figure 14. Timing During a DRAM Write During Bus Grant Mode
CLKIN
ASI
D<31:0>
RDWR
BGRNT
AS
BREQ
Page MISS-HITPage Hit
Idle
VALID VALID
Row Column Column
I/O Device output
ADR<27:2>ADR1/(BE2)
CS4
RDYOUT
RASx
CASx
BE<3:0>
I/O Device output
CPU Driving
External Circuit driving
Floating
DWEx
Pullup required, as BE2 is input mode
MB86831
Fujitsu Microelectronics, Inc. 39
Figure 15. 8-Bit Bus Mode (1 Wait State) for Read Only
Figure 16. 8-Bit Bus Mode (1 Wait State) for Word Write
CLKIN
AS
ADR<27:2>
BE<3:0>
READY
DATA
0000 0001 0010 0011
Byte 0 Byte 1 Byte 2 Byte 3
RDWR
CLKIN
AS
ADR<27:2>
BE<3:0>
READY
0011 0010 0001 0000
RDWR
DATA BYTE3 BYTE2 BYTE1 BYTE0
SPARClite Series 32-Bit RISC Embedded Processor
40 Fujitsu Microelectronics, Inc.
Figure 17. 16-Bit Bus Mode (1 Wait State) for Read only
Figure 18. 16-Bit Bus Mode (1 Wait State) for Word Write
CLKIN
AS
ADR<27:2>
BE<3:0>
READY
DATA
0000 0010 0000 0010
HW0 HW1 HW0 HW1
RDWR
CLKIN
AS
ADR<27:2>
BE<3:0>
READY
DATA
0000 0010 0000 0010
HW0 HW1 HW0 HW1
RDWR
MB86831
Fujitsu Microelectronics, Inc. 41
Figure 19. DRAM Controller Enabled, DRAM Burst Disabled, 32-Bit Page Mode, Read/Write Timing
Figure 20. DRAM Controller Enabled, DRAM Burst Enabled, 32-Bit Page Mode, Read/Write Timing
CLKIN
AS
RDWR
ADR<13:2>
RASx
CASx
RDYOUT
DWEx
D<31:0>
DOE
SAMEPAGE
Read (Miss) Read (Hit) Write (Hit) Read (Miss)Read (Hit)
TCAS
TRP +1CLK
Column Column Column Column ColumnRow Row
Note: During reads there are instruction fetches and loads. If the pre-fetch buffer is operating during instruction fetches, idle cycles may not open between reads.
CLKIN
AS
RDWR
ADR<13:2>
RASx
CASx
RDYOUT
DWEx
D<31:0>
DOE
SAMEPAGE
Read (Hit) Write (Hit) Write (Miss) Read (Hit)
TCAS
TRP +1CLK
Column Column Column Column ColumnRow ColumnColumn
SPARClite Series 32-Bit RISC Embedded Processor
42 Fujitsu Microelectronics, Inc.
Figure 21. DRAM Controller Enabled, DRAM Burst Enabled, 32-Bit EDO, Read Timing
Figure 22. DRAM Controller Enabled, DRAM Burst Enabled, 32-Bit EDO, Read/Write Timing
CLKIN
AS
RDWR
ADR<13:2>
RASx
CASx
RDYOUT
DWEx
D<31:0>
DOE
SAMEPAGE
Read (Miss) Read (Miss)Read (Hit)
TRP +1CLK
ColumnColumn Column Column ColumnRow ColumnColumn Column Row
CLKIN
AS
RDWR
ADR<13:2>
RASx
CASx
RDYOUT
DWEx
D<31:0>
DOE
SAMEPAGE
Write (Miss)Write (Hit)Read (Hit)
TRP +1CLK
ColumnColumn Column Column Row ColumnColumn Column
Read (Hit)
Column Column Column
TCAS
Note: If EDO DRAM is selected, CAS width during reads is 1 clock, and CAS width during writes is the TCAS setting.
MB86831
Fujitsu Microelectronics, Inc. 43
.
Figure 23. DRAM Controller Enabled, DRAM Burst Disabled, 16-Bit Page Mode, Read/Write Timing
Figure 24. DRAM Controller Enabled, DRAM Burst Enabled, 16-Bit Page Mode Read Timing 1
CLKIN
AS
RDWR
ADR<13:2>
RASx
CAS3,-2
RDYOUT
DWEx
D<15:0>
DOE
SAMEPAGE
Write (Hit)Read (Miss)
TRP +1CLK
Column ColumnColumn ColumnRow Column Column
Read (Hit)
Row
TCAS
Note: During reads there are instruction fetches and loads. If the pre-fetch buffer is working during instruction fetches, idle cycles may not open between reads.
CLKIN
AS
RDWR
ADR<13:2>
RASx
CAS3,-2
RDYOUT
DWEx
D<15:0>
DOE
SAMEPAGE
Read (Miss)
TRP +1CLK
Column ColumnColumn ColumnRow
TCAS
Column ColumnColumn Column
“H”
“H”
SPARClite Series 32-Bit RISC Embedded Processor
44 Fujitsu Microelectronics, Inc.
Figure 25. DRAM Controller Enabled, DRAM Burst Enabled, 16-Bit EDO, Read Timing 1
Figure 26. DRAM Controller Enabled, DRAM Burst Enabled, 16-Bit EDO, Read Timing 2
CLKIN
AS
RDWR
ADR<13:2>
RASx
CAS3,-2
RDYOUT
DWEx
D<15:0>
DOE
SAMEPAGE
Read (Miss)
TRP +1CLK
Column ColumnColumn ColumnRow Column ColumnColumn Column ColumnRow
Read (Miss)
“H”
“H”
CLKIN
AS
RDWR
ADR<13:2>
RASx
CAS3,-2
RDYOUT
DWEx
D<15:0>
DOE
SAMEPAGE
Read (Hit)
Column ColumnColumn Column Column ColumnColumn Column
Read (Hit)
Column ColumnColumn Column Column Column
“H”
“H”
“L”
MB86831
Fujitsu Microelectronics, Inc. 45
Figure 27. DRAM Controller Enabled, DRAM Burst Enabled, 16-Bit EDO, Write Timing
Figure 28. Example of PBREQ Timing
CLKIN
AS
RDWR
ADR<13:2>
RASx
CAS3,-2
RDYOUT
DWEx
D<15:0>
DOE
SAMEPAGE
Write (Miss)
TRP +1CLK
Column ColumnRow Column Column ColumnColumn ColumnRow
Write (Miss) Write (Hit) Write (Hit)
TCAS
“H”
CLKIN
AS
PBREQ
ADR<27:2>
READY
DATA
BREQ
BGRNT
t1 t2 t4t3
SPARClite Series 32-Bit RISC Embedded Processor
46 Fujitsu Microelectronics, Inc.
Electrical CharacteristicsTable 17. Absolute Maximum Ratings VSS=0V
Notes: 1. Stresses above those listed under Absolute Rating may cause permanent damage to the device. This is a stress rating only and functional operation of the device at these or any other condition above those indicated in the operation section of this specification is not implied. Exposure to Absolute Maxi-mum Ratings conditions for extended periods may affect device reliability.
Recommended Connections:1. Power and ground connections must be made to multiple VDD and VSS pins. Every MB86831 based circuit board should include power (VDD) and ground
(VSS) planes for power distribution. Every VDD pin must be connected to the power plane, and every VSS pin must be connected to the ground plane. Pins identified as “N.C.” must not be connected in the system.
2. Liberal decoupling capacitance should be placed near the MB86831. The processor can cause transient power surges when multiple output buffers tran-sition, particularly when connected to large capacitive loads.
3. Low inductance capacitors and interconnections are recommended for best high frequency electrical performance. Inductance can be reduced by short-ening the board traces between the processor and decoupling capacitors as much as possible. Capacitors specifically designed for the QFP package offer the least inductance.
4. For reliable operation, alternate bus masters must drive any pins that are floated by the MB86831 when it has granted the bus, in particular the LOCK, ADR<27:2>, ASI<3:0>, BE<3:0>, D<31:0>, AS and RDWR signals must be driven by alternate bus masters. These pins are normally driven by the processor during active and idle bus states and don’t require external pullups (except for BE2, which becomes an address input). N.C. pins must always remain unconnected.
Table 18. Recommended Operating Conditions
• The MB86831 can be used with a 5V system or a 3.3V system interface:
5V System Interface IO_VDD = 5V and VDD = 3.3V Two Power Supplies3.3V System Interface IO_VDD = VDD = 3.3V Single Power Supply
Symbol Rating Min. Max. Units
IO_VDD Power Supply Voltage (I/O) -0.5 6 V
VDD Power Supply Voltage (Core) -0.5 4 V
V1 Input Voltage -0.5 IO_VDD+0.5 V
TSTG Storage Ambient Temperature -55 125 °C
TBIAS Temperature during Bias 0 70 °C
- Overshoot IO_VDD+1.0V or less (50ns or less)
- Undershoot VSS-1.0V or less (50ns or less)
Symbol Rating Min. Typ. Max. Units
IO_VDD Power Supply Voltage (I/O = 5V) 4.75 5.0 5.25 V
Power Supply Voltage (I/O = 3.3V) 3.0 3.3 3.6 V
VDD Power Supply Voltage (Core) 3.0 3.3 3.6 V
VIL Low Level Input Voltage 0 - VDD×0.25 V
VIH High Level Input Voltage VDD×0.65 - IO_VDD V
Topr Operating Temperature 0 25 70 °C
MB86831
Fujitsu Microelectronics, Inc. 47
• Because 5V inputs cannot be accepted using the 3.3V sys-tem interface, all input signals must be driven at 3.3V logic levels.
• With the 5V system interface, outputs are driven with 5V logic levels. The input is defined by 3.3V logic levels, but inputs greater than 3.3V can also be tolerated. The CPU and core logic power supply current is the same for both 3.3V and 5V operation.
• In the 5V system interface, two power supplies are required. While there are basically no restrictions on power supply sequencing, the following sequence is recom-mended:
At power-on VDD IO_VDD SignalAt power-off Signal IO_VDD VDD
DC Characteristics5V Interface IO_VDD = 5V ± 5% VDD = 3.3V ± 0.3V, VSS = 0V, Ta = 0~70°C
3.3V Interface IO_VDD = VDD = 3.3V ± 0.3V, VSS = 0V, Ta = 0~70°C
Symbol Rating Conditions Min. Typ. Max. Units
VIL Low Level Input Voltage 0 - VDD×0.25 V
VIH High Level Input Voltage VDD×0.65 - IO_VDD V
VOL Low Level Output Voltage IOL = 4 mA 0 - 0.4 V
VOH High Level Output Voltage IOH = -4 mA IO_VDD-0.5 - IO_VDD V
ILI Input Leak Current VIN = 0 or IO_VDD -10 - 10 µA
ILZ Floating Output Leakage Current VOUT = 0 or IO_VDD -10 - 10 µA
IDD I/O Power Supply Current (IO_VDD) 33 MHz No Load - 40 50 mA
IDD CPU/Core Power Supply Current (VDD) 66 MHz - 150 187.5 mA
IDD CPU/Core Power Supply Current (VDD) 80 MHz - 200 250 mA
ISLEEP CPU/Core Power Supply Current (VDD) in Sleep Mode 66 MHz - 10 - mA
PD Power Consumption (IO_VDD+VDD) 66 MHz No Load - 950 1188 mW
CPIN Pin CapacityIO_VDD = V1 = 0f = 1 MHz
- - 16 pF
Symbol Rating Conditions Min. Typ. Max. Units
VIL Low Level Input Voltage 0 - VDD×0.25 V
VIH High Level Input Voltage VDD×0.65 - IO_VDD V
VOL Low Level Output Voltage IOL = 2.0 mA 0 - 0.4 V
VOH High Level Output Voltage IOH = -2.0 mA IO_VDD-0.5 - IO_VDD V
ILI Input Leak Current VIN = 0 or IO_VDD -10 - 10 µA
ILZ Floating Output Leakage Current VOUT=0 or IO_VDD -10 - 10 µA
IDD I/O Power Supply Current (IO_VDD) 33 MHz No Load - 30 37.5 mA
IDD I/O Power Supply Current (IO_VDD) 40 MHz No Load - 36 45 mA
PD Power Consumption (IO_VDD+VDD) 66 MHz No Load - 594 743 mW
CPIN Pin CapacityIO_VDD = V1 = 0f = 1 MHz
- - 16 pF
SPARClite Series 32-Bit RISC Embedded Processor
48 Fujitsu Microelectronics, Inc.
Alternating current characteristics are all regulated by CLKIN (BIU clock), and they do not depend on the CPU internal operating frequency.
AC CharacteristicsVDD = 3.3V ± 0.3V, VSS = 0V, Ta = 0~70°C, P=Period (CLKIN period)
Symbol Parameter DescriptionIO_VDD = 5V + 5% IO_VDD = 3.3V + 0.3V
UnitsMin. Max. Min. Max.
t1 CLKIN period 30 100 30 100 ns
t2 CLKIN high time 10 - 10 - ns
t3 CLKIN low time 10 - 10 - ns
t4 CLKIN rise time - 3 - 3 ns
t5 CLKIN fall time - 3 - 3 ns
t6
D<31:0>Output valid delay - 20 - 20 ns
Output hold 2 - 2 - ns
ADR<27:2>Output valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
BE<3:0>Output valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
ASI<3:0>Output valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
CS<5:0>Output valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
t7
RDWROutput valid delay - 18 - 19 ns
Output hold 2 - 2 - ns
LOCKOutput valid delay - 18 - 19 ns
Output hold 2 - 2 - ns
ASOutput valid delay - 18 - 19 ns
Output hold 2 - 2 - ns
TIMER_OVFOutput valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
BGRNTOutput valid delay - 18 - 19 ns
Output hold 2 - 2 - ns
PBREQOutput valid delay - 18 - 19 ns
Output hold 2 - 2 - ns
RDYOUT(Internal READY mode)
Output valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
RDYOUT(External READY mode)
Output valid delay - 15 - 15 ns
Output hold 2 - 2 - ns
ERROROutput valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
PDOWNOutput valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
MB86831
Fujitsu Microelectronics, Inc. 49
AC Characteristics (continued)
Symbol Parameter DescriptionIO_VDD = 5V + 5% IO_VDD = 3.3V + 0.3V
UnitsMin. Max. Min. Max.
t8
SAMEPAGEOutput valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
TIMER_OVFOutput valid delay - 20 - 21 ns
Output hold 2 - 2 - ns
BMREQOutput valid delay - 18 - 19 ns
Output hold 2 - 2 - ns
t9
RAS<3:0>Output valid delay - 15 - 15 ns
Output hold 2 - 2 - ns
CAS<3:0>Output valid delay - 15 - 15 ns
Output hold 2 - 2 - ns
DWE<3:0>Output valid delay - 15 - 15 ns
Output hold 2 - 2 - ns
DOEOutput valid delay - 15 - 15 ns
Output hold 2 - 2 - ns
t10
READYInput setup time 14 - 14 - ns
Input hold time 2 - 2 - ns
MEXCInput setup time 14 - 14 - ns
Input hold time 2 - 2 - ns
BREQInput setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
BMACKInput setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
RDWRInput setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
ASInput setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
t11
ASI<3:0>Input setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
ADR<27:2>Input setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
BE2Input setup time 12 - 12 - ns
Input hold time 2 - 2 - ns
t12 D<31:0>Input setup time 14 - 14 - ns
Input hold time 2 - 2 - ns
IRL<3:0>Input setup time Asynchronous Asynchronous
Input hold time Asynchronous Asynchronous
WKUPInput setup time Asynchronous Asynchronous
Input hold time Asynchronous Asynchronous
SPARClite Series 32-Bit RISC Embedded Processor
50 Fujitsu Microelectronics, Inc.
AC Characteristics (continued)
Notes: 1. In the absence of documentation to the contrary, all parameters are valid within the specified temperature and power supply ranges. These specifications are subject to change for performance improvement.
2. All voltage values have the GND (VSS = 0V) as their reference. Timing measurement reference points are 1.5V, input level from 0.4V to 2.4V, and input rise and fall times are 2ns or less.
3. Do not short multiple output pins for more than 1 second. External output load capacity is 30 pF.4. Pins which are specified for asynchronous input, and standards other than RDYOUT during external READY mode, are referenced from the external clock
(CLKIN) rising edge.5. RDYOUT during external READY mode is referenced from the READY input.6. A minimum of four CLKIN Cycles are required for reset initialization. Also, 4,000 CLKIN cycles are required for PLL oscillation stabilization time.
For example, when CLKIN is 33 MHz (30ns).
Figure 29. Reset Timing
Figure 30. Clock Timing
Symbol Parameter DescriptionIO_VDD = 5V + 5% IO_VDD = 3.3V + 0.3V
UnitsMin. Max. Min. Max.
IRQ<15:8> Input setup time Asynchronous Asynchronous ns
IRQ<15:8> Input hold time Asynchronous Asynchronous ns
IRQ<15:8> Input High level duration 2×P+10 - 2×P+10 - ns
IRQ<15:8> Input Low level duration 2×P+10 - 2×P+10 - ns
CLKIN
ADDR
RESET
4 Cycle Minimum
0x0000 0000
*1
*1. CLKIN must be stable before RESET is deasserted.*2. When RESET hold time is met.
3 Cycles *2
t1
t5t2 t3
t4
2.8v 2.8v 2.8v
0.8v 0.8v1.5v 1.5vCLKIN
MB86831
Fujitsu Microelectronics, Inc. 51
Figure 31. I/O Output Timing
CLKIN
AS,RDWR,LOCK,
D<31:0>,ADR<27:2>,
1.5v
1.5v
t6h
1.5v
t7h
1.5v
t8h
RAS<3:0>,CAS<3:0>, 1.5v
t9h
t6d
t7d
t8d
t9d
BE<3:0>,ASI<3:0>,CS<5:0>
BGRNT,PBREQ,RDYOUT,
SAME_PAGE,TIMER_OVF,BMREQ
ERROR,PDOWN
DWE<3:0>,DOE
SPARClite Series 32-Bit RISC Embedded Processor
52 Fujitsu Microelectronics, Inc.
Figure 32. I/O Input Timing
1.5v
1.5v
t12s t12h
1.5v
t11s t11h
CLKIN
D<31:0>
NOTE: s=setup, h=hold
ADR<27:2>
RDWR,READY,MEXC,1.5vBMACK,AS,BREQ
t10d t10h
ASI<3:0> 1.5v
t11s t11h
ADR1/BE2
MB86831
Fujitsu Microelectronics, Inc. 53
Exterior Package DrawingPlastic. SQFP, 176 Pin EIAJ Code: *QFP176-P-2424-1
(FPT-176P-M01)
Lead Pitch 0.50mm
Package Width ×Package Length
24 × 24mm
Lead Profile Gull Wing
Sealing Method Plastic Mold
(0.20±0.10)
.008±.0040.50(.0197)
TYP
Dimensions in inches (millimeters)
176-LEAD PLASTIC QFP PACKAGE(CASE No.: FPT-176P-M01)
Details of “A” part
0.08(.003) M
21.50(.846)REF
0(0) MIN
(STAND OFF)
3.85(.152) MAX
(INSTALLED HEIGHT)
25.60(1.008)NOM
0.25(.010)
0.15(.006)
MAX
“A”
0.10(.004)
45
88
89132
133
176 44
1LEAD No.
INDEX
(0.125±0.05)
.005±.002
“B”
0.40(.016)
MAX
.0.50±0.20
(020±.008)
Details of “B” part
0.20(.008)
26.60±.0.20 (1.047±0.008)SQ
24.00±0.10 (.945±0.004)SQ
0~10°
Copyright© 1998 Fujitsu Microelectronics Ltd. andFujitsu Microelectronics Inc., USA. All Rights Reserved.
All company and product names are trademarks or registered trademarks of their respective owners.
The information contained in this document has been carefully checked and is believed to be reliable. However, Fujitsu Microelectronics, Inc. assumes no responsibility for inaccuracies.
The information in this document does not convey any license under the copyrights, patent rights, or trademarks claimed and owned by Fujitsu Limited, its subsidiaries, or Fujitsu Microelectronics, Inc.
Fujitsu Microelectronics, Inc. reserves the right to change products or specifications without notice.
No part of this publication may be copied or reproduced in any form, or by any means, or transferred to any third party without prior written consent of Fujitsu Microelectronics, Inc.
Printed in U.S.A.
EC-DS-20386-8/98
FUJITSU MICROELECTRONICS, INC.
Corporate Headquarters3545 North First Street, San Jose, California 95134-1804Tel: (800) 866-8608 Fax: (408) 922-9179E-mail: [email protected] Internet: http://www.fujitsumicro.com